Scribal Habits of Codex Sinaiticus 9781463211592

Citation preview

Scribal Habits of Codex Sinaiticus

Texts and Studies

5 Series Editor Hugh Houghton

Editorial Board Jeff W. Childers Christina M. Kreinecker Alison G. Salvesen Peter John Williams

Text and Studies is a series of monographs devoted to the study of Biblical and Patristic texts. Maintaining the highest scholarly standards, the series includes critical editions, studies of primary sources, and analyses of textual traditions.

Scribal Habits of Codex Sinaiticus

Dirk Jongkind

9

34 2013

Gorgias Press LLC, 954 River Road, Piscataway, NJ, 08854, USA www.gorgiaspress.com Copyright © 2013 by Gorgias Press LLC

All rights reserved under International and Pan-American Copyright Conventions. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise without the prior written permission of Gorgias Press LLC. 2013

‫ܝ‬

9

ISBN 978-1-59333-422-2

Printed in the United States of America

ISSN 1935-6927

CONTENTS Contents ................................................................................................................. vii Preface ................................................................................................................... xiii Abbreviations ...................................................................................................... xvii Introduction............................................................................................................. 1 1 History of Research....................................................................................... 5 Publication of the Codex.............................................................................. 5 Scribes and Correctors .................................................................................. 9 Date and Provenance .................................................................................. 18 Dictation........................................................................................................ 21 Text ............................................................................................................... 24 History of Research and This Study ......................................................... 26 2 The Non-textual Characteristics of Scribal Behaviour in Codex Sinaiticus ......................................................................................................... 29 Introduction.................................................................................................. 29 The Constitutive Elements......................................................................... 30 Sheet and quire....................................................................................... 30 Folio and page ........................................................................................ 34 Column and line..................................................................................... 36 The use of red ink.................................................................................. 37 Scribes............................................................................................................ 39 Overview of the work of the scribes .................................................. 39 Interaction between scribes A and D ................................................. 41 Change of scribe .............................................................................. 41 Correction of each other’s work—cancel leaves ........................ 44 Correction of each other’s work—corrections on the text....... 46 Division of tasks and space restrictions: Revelation, Barnabas, Hermas ......................................................................... 48 Supplementary apparatus: superscriptions and running titles ................................................................................................. 51 Conclusions................................................................................................... 55 vii

Contents

3

The scribes .............................................................................................. 55 The production of the manuscript...................................................... 57 Concluding observations ...................................................................... 59 Nomina Sacra, Ligatures, Itacisms, Text-divisions ................................... 61 Introduction.................................................................................................. 61 Nomina Sacra ............................................................................................... 62 The forms of the nomina sacra............................................................... 64 Relative frequency of the contracted versus the uncontracted forms ...................................................................... 67 Differences between the scribes.......................................................... 74 The data............................................................................................. 74 The individual patterns for each scribe .............................................. 79 Scribe A ............................................................................................. 79 Scribe D............................................................................................. 80 Scribe B.............................................................................................. 81 Conclusions on the use of nomina sacra in the analysed sample ............................................................................................. 82 Ligatures ........................................................................................................ 84 Combining letters................................................................................... 84 mou and pro(s) ligatures................................................................. 87 The kai-ligature.................................................................................... 88 Conclusions on the use of ligatures .................................................... 89 Orthographic Patterns................................................................................. 90 Introduction............................................................................................ 90 Itacisms per scribe ................................................................................. 91 Itacisms in proper nouns ...................................................................... 93 Conclusions on itacisms........................................................................ 94 Paragraphing................................................................................................. 95 Introduction............................................................................................ 95 Methods to indicate paragraphs .......................................................... 95 Frequency of new paragraphs.............................................................. 97 Differences within the work of a scribe............................................. 98 The quality of paragraph breaks ........................................................ 100 Scribe B in the prophets ............................................................... 100 Scribe A in the Gospels ................................................................ 102 Scribe A in 4 Maccabees ............................................................... 106 Scribe A in Romans....................................................................... 107 Concluding observations on the use of paragraphs ....................... 108 Eusebian Apparatus................................................................................... 109 Introduction.......................................................................................... 109

Contents

4

The positioning of the Eusebian apparatus in the margin ............ 112 Scribal customs and clear scribal errors ..................................... 112 Substantial deviations in placement ............................................ 115 Deviating section and table numbers ............................................... 116 Discussion of deviating placements and deviating numbers ........ 117 Concluding observations on the Eusebian apparatus.................... 119 Numbering Systems and Kephalia ............................................................ 120 The section numbering in Acts ......................................................... 121 Kephalia in Acts ..................................................................................... 122 The section numbering in Ecclesiastes and Song of Songs .......... 124 Concluding observations on the numbering systems and kephalia .......................................................................................... 126 Conclusions................................................................................................. 127 Singular Readings and the Individual Scribal Patterns of Scribes A and D ........................................................................................................... 131 Introduction................................................................................................ 131 Methodological Considerations............................................................... 131 Past scholarship and the study of singular readings ....................... 131 The nineteenth century: Tischendorf and Westcott-Hort ...... 131 The twentieth century: Colwell and Royse ................................ 134 The method of studying singular readings and Codex Sinaiticus ........................................................................................ 140 System of classification of singular readings ................................... 142 Singular Readings and Corrections in 1 Chronicles ............................. 144 Introduction.......................................................................................... 144 The text and the correctors ................................................................ 145 Discussion of the evidence................................................................. 147 Orthography ................................................................................... 147 Nonsense words............................................................................. 150 Leaps from the same to the same ............................................... 151 Addition/omission of verba minora........................................... 151 Harmonisations .............................................................................. 152 Nonsense in context, substitutions, and transpositions .......... 156 Add and omit words and clauses................................................. 157 Correction strategy .............................................................................. 159 Scribe created readings that are not singular readings. .................. 160 Final reflections on 1 Chronicles....................................................... 163 Scribes D and A in Psalms ....................................................................... 164 Introduction.......................................................................................... 164 Method .................................................................................................. 165

Contents Corrections by the scribes .................................................................. 167 Discussion of the singular readings .................................................. 169 Orthography ................................................................................... 170 Nonsense word forms................................................................... 172 Leaps ................................................................................................ 175 Add and omit verba minora......................................................... 179 Harmonisations .............................................................................. 182 Editorial readings ........................................................................... 188 Nonsense meanings....................................................................... 189 Substitutions ................................................................................... 191 Transpositions ................................................................................ 196 Addition and omission of words and phrases........................... 198 Major rewritings ............................................................................. 199 Conclusions........................................................................................... 200 Scribes D and A in Paul............................................................................ 202 Introduction.......................................................................................... 202 Discussion of the singular readings per category............................ 204 Orthography ................................................................................... 204 Nonsense word forms................................................................... 205 Leaps from the same to the same ............................................... 206 Add and omit verba minora......................................................... 208 Harmonisations .............................................................................. 210 Editorial readings ........................................................................... 213 Nonsense meanings in context.................................................... 214 Substitutions ................................................................................... 214 Transpositions ................................................................................ 216 Add and omit words and clauses................................................. 217 Major rewritings ............................................................................. 218 Conclusions........................................................................................... 219 Scribes D and A in Luke........................................................................... 221 Introduction.......................................................................................... 221 Discussion of the singular readings per category............................ 223 Orthography ................................................................................... 223 Nonsense words............................................................................. 225 Leaps from the same to the same ............................................... 227 Add and omit verba minora......................................................... 228 Harmonisations .............................................................................. 229 Editorial readings ........................................................................... 232 Nonsense meanings....................................................................... 233 Substitutions ................................................................................... 234

Contents Transpositions ................................................................................ 236 Add and omit words and clauses................................................. 237 Rewritings........................................................................................ 240 Conclusions........................................................................................... 240 Conclusions................................................................................................. 241 On using singular readings to study scribal habits ......................... 241 The individual scribal patterns for scribes D and A....................... 242 On scribal habits in general................................................................ 245 5 Final Reflections......................................................................................... 247 Scribal Behaviour ....................................................................................... 247 The Codex and the Exemplar.................................................................. 250 Dictation Theory........................................................................................ 250 Date and Provenance ................................................................................ 252 Appendix I: Tables and Graphs to Chapter 3................................................ 257 Appendix II: List of the Substantial Differences of Position in the Eusebian Apparatus between Sinaiticus and NA27............................... 263 Matthew................................................................................................. 263 Mark ....................................................................................................... 268 Luke ....................................................................................................... 275 John ....................................................................................................... 276 Appendix III: Places Where the Eusebian Apparatus is Missing................ 281 Appendix IV: Deviating and Missing Table Numbers in the Eusebian Apparatus .................................................................................................... 283 Appendix V: Uncorrected Singular Readings in 1 Chronicles..................... 287 Appendix VI: The Corrected Readings of Psalms ........................................ 297 Bibliography......................................................................................................... 305 Index of Modern Authors ................................................................................. 313 Index of Scriptural Passages.............................................................................. 314

PREFACE In general, a doctoral thesis is tough reading and I am afraid this work is no exception. The reader will be confronted with long lists of readings and endless tables summarising extensive counts, all of which are presented in a prose that betrays the fact that the author is not a native English speaker. I do apologise for these flaws but gain some consolation from the notion that I never intended this work to fulfil any literary aspirations I might have. Rather, I have sought to make a contribution to the field of textual criticism of the Greek Bible, with a noticeable bias towards the New Testament side. I have aimed to present as much of the data as possible, which does not make for entertaining reading but does help the cause of verifiable scholarship. Despite checking, re-checking, and many rounds of corrections, it is almost inevitable that I will have made mistakes. I had to deal with a large amount of data and made many individual judgments on variants. Without doubt the reader will disagree with a number of my decisions. I do hope, though, that this will not affect the argument I try to make. This study is a lightly revised version of the thesis I presented to the Faculty of Divinity in Cambridge in 2005. In the more than four years that it took to write the original thesis, many people have played a role in ensuring that it would be completed. Among them are, without doubt, my friends and fellow-labourers at Tyndale House. The warden, Dr. Bruce W. Winter, was instrumental in getting me to Cambridge in the first place, and throughout the years he has provided warm and welcome encouragement. Also the other members of staff at Tyndale House have been tremendous throughout my time with them and it is a privilege to have joined their ranks. The Faculty of Divinity has been a most stimulating place to undertake the study of Codex Sinaiticus. It has a long-standing tradition of primary textual research which is still very much alive. Not only did the faculty take an active interest in my research by inviting me to read a paper to the Senior New Testament Seminar, but they also gave me plenty of opportunity to xiii

Preface connect with a number of leading experts in the field. I wish to thank Professor Graham Stanton for the role he has played in all this. My supervisor, Dr. Peter Head, has been a tremendous inspiration through his advice, guidance, and knowledge of the field. I feel privileged to have worked under someone who took a sincere interest in the whole project from the beginning to the very end. My examiners, Professor David Parker and Professor J.K. Elliott, not only encouraged me to seek publication of my thesis, but also pointed out a fair number of possible improvements. I thank them for their support. A special word of thanks is due to the Manuscript Department of the British Library. The Head of Medieval and Earlier Manuscripts, Dr. Scot McKendrick, ensured that during the writing of the thesis I gained access to the manuscript in order to verify information in Milne and Skeat’s Scribes and Correctors and to answer some of the questions raised in the course of my research. After the thesis was submitted for examination, but before it was accepted for publication, I had the privilege of being employed by the British Library as curator in the Codex Sinaiticus Digitisation Project. This longer exposure to the original manuscript lead me to change some of the details of the original doctoral thesis and also raised new questions which, I hope, will be answered during the course of the various studies undertaken in connection with the Digitisation Project. I am convinced that the present study will help others to see more in the manuscript, especially now that we can expect an exciting, new digital facsimile of Sinaiticus. My church family at North Arbury Chapel, Cambridge, have surrounded me and my family with all their love, support, and prayers. They helped me balance the academic study of textual transmission and the practice of the Christian life. I thank them for giving me true Christian freedom. I am blessed with a wonderful family. Without my children, Jonathan, Reuven, Hannah, Elisabeth, Joshua, and Sarah Grace, the writing of this work would have taken much longer. It is good to know that one’s harshest critics are those who love you most. My wife Marion has been a constant source of energy and I am grateful for having such a companion. Tischendorf, whose name is intrinsically linked with Codex Sinaiticus, wrote the following words: “Providence has given to our age, in which attacks on Christianity are so common, the Sinaitic Bible, to be to us a full and clear light as to what is the real text of God’s Word written, and to assist us in defending the truth by establishing its authentic form.”1 Unlike C. Tischendorf, When Were Our Gospels Written? (London: Religious Tract Society, 1896), 34. 1

Preface Tischendorf, I am not convinced that Codex Sinaiticus is a “full and clear light” in the textual criticism of the Greek Bible. Nevertheless, the manuscript does assist us in establishing the authentic form of Scripture. I am grateful to Him who is the God of Scripture for allowing me to play my own, minor, part in the study of that Book. Dirk Jongkind Autumn 2006

ABBREVIATIONS Abbreviations have been avoided as much as possible, but when used they comply with The SBL Handbook of Style, edited by Patrick H. Alexander et al. (Peabody, Mass.: Hendrickson, 1999). Some frequently cited works are referenced by the following abbreviations: Brooke-McLean A.E. Brooke, Norman McLean, and Henry St. John Thackeray. The Old Testament in Greek: according to the Text of Codex Vaticanus, Supplemented from other Uncial Manuscripts, with a Critical Apparatus Containing the Variants of the Chief Ancient Authorities for the Text of the Septuagint. Vol. II: The Later Historical Books, part III: I and II Chronicles. Cambridge: Cambridge University Press, 1932. Gignac I, II Francis Thomas Gignac. A Grammar of the Greek Papyri of the Roman and Byzantine Periods. Vol. I Phonology and Vol. II - Morphology. Testi e documenti per lo studio dell’antichità 55. Milano: Istituto editoriale cisalpino - La Goliardica, 1976– 81. Milne-Skeat H.J.M. Milne and T.C. Skeat, with contributions by Douglas Cockerell. Scribes and Correctors of the Codex Sinaiticus. London: British Museum, 1938. Thackeray Henry St. John Thackeray. A Grammar of the Old Testament in Greek. Cambridge: University Press, 1909. Westcott-Hort B.F. Westcott and F.J.A. Hort. The New Testament in the Original Greek: Introduction, Appendix. 2nd ed. London: Macmillan, 1896.

xvii

INTRODUCTION The following work contains a series of studies of the Codex Sinaiticus, further referred to as Sinaiticus. This name is used to describe all the parts of the Greek codex that are now preserved in four different places (London, Leipzig, St Petersburg, and St Catharine’s monastery on Mt Sinai), and of which the first published part was originally given the name Codex FridericoAugustanus. Though I trust the results of this study will be sufficient justification for undertaking the project, some remarks on our aim may be in order. When, in the autumn of 2004, I told an eminent text critic that I was writing a thesis on Sinaiticus, her initial reaction was, “Is there anything left to study since Milne and Skeat’s work?” In the subsequent conversation I was able to explain what I was doing, but the response serves the point that a general feeling may exist that, since Milne and Skeat’s monograph of 1938, little remains to be done. In this study we will focus on scribal behaviour. This broad term includes everything from the way the scribes prepared their writing material to the way in which they handled the text. Fortunately, Milne and Skeat have done much of the codicological and palaeographical work. In our study we will go one step further and use their results to form a better understanding of all the scribal activity present in Sinaiticus. Most of our studies limit themselves to the boundaries of the codex; we do not aim to undertake a large comparative study of the wider tradition or the contemporary manuscripts (one of the few exceptions is our study of the Eusebian apparatus in Chapter 3). No separate study was made of the bookhand that the scribes used, the biblical Majuscule; a thorough monograph has already been published on this subject by an expert palaeographer.1 Our stated aim also excludes the later history of the manuscript. What happened to the manuscript after the scribes finished working on it, such as the corrections made on the manuscript in later centuries and the controversy surrounding its so-called discovery by Constantin Tischendorf in the 19th century, lie beyond the boundary of our 1

G. Cavallo, Ricerche sulla Maiuscola Biblica (Firenze: Le Monnier, 1967). 1

2

Introduction

study. For the sake of completeness, however, we have included discussions of the scribal hands of the so-called C correctors, who did their work only a few centuries after the manuscript was written. Their contributions must be subtracted from the current shape of the manuscript in order to arrive at the manuscript that left the hands of the scribes. Most of the work for this study has made use of the facsimile editions of the Old and New Testament by Helen and Kirsopp Lake, supported by Tischendorf’s notes as published in his text editions of 1862 and 1863 and by the descriptions and lists contained in Scribes and Correctors of the Codex Sinaiticus, by H.J.M. Milne and T.C. Skeat. Without the detailed description by Milne and Skeat this study would have been unthinkable; they are the giants on whose shoulders we stand. To indicate a particular passage in the manuscript, a rather extensive means of referencing is used. A particular location is normally indicated as, e.g., Matt 22:29, folio 74.5 (NT 13), line 6.25. First the canonical reference is given, then the number of the quire (here quire 74) followed by the folio.2 Thus folio 74.5 means the fifth folio of quire 74. Between brackets the folio number of Lake’s facsimile is indicated, referring either to the Old Testament or New Testament volume. Thus, folio 13 of the New Testament is cited in the example above. Finally the column and line of the specific reference are given, where the column number is counted from the first column of the recto page. For text set in prose, this means that columns 1–4 are on the recto side, and 5–8 on the verso; in the books of the LXX that are set in the poetic layout, columns 1–2 are on the recto, 3–4 on the verso. In our example, line 6.25 indicates the 25th line of column 6, which is the second column on the verso side of the folio. Incidentally, when the reader comes across expressions such as 26.5 folios, they indicate the quantity of folios (here, 26 and a half) and not a combination of quires and folios. In chapters 2–4, the reader will occasionally find text from Sinaiticus laid out according to the original line and paragraph division. However, one should realise that to a certain degree these texts have been edited: divisions between words have been added, some ligatures or contractions may have been resolved, and the crossbar “ ) ” —indicating a final nu—may have been written out in full. One final apology is necessary. In this study, reference is made to the scribes using the masculine pronoun: “he wrote”, “he made a mistake”, etc. In accordance with Milne-Skeat, we use the quire numbering visible in the top right corner of the first page of each quire. 2

Introduction

3

I am aware that this might give the impression that the author believes that the scribes were male. This is not the case; I simply do not know. There is literary evidence of female scribes in the earliest centuries.3 The masculine pronouns should be read in a generic rather than a gender-specific way.

Eusebius mentions young women in the ‘production team’ around Origen (Hist. eccl. 6.23.2). 3

1 HISTORY OF RESEARCH

PUBLICATION OF THE CODEX Constantin von Tischendorf brought the first part of the codex to the West in January 1845 and published the 43 leaves the next year under the name of Codex Friderico-Augustanus.1 He remained vague about the origin of these leaves and did not mention that he had only been allowed to take part of what he had seen in St Catherine’s monastery at the foot of Mt Sinai. Apparently Tischendorf had only seen the prose sections of the OT, totalling 130 leaves, for he later tells us that he brought one third of the leaves to Leipzig.2 The Codex Friderico-Augustanus contains part of 1 Chronicles and the end of 2 Esdras, Esther and the beginning of Tobit, the last part of Jeremiah and the first part of Lamentations.3 During his first visit to Mt Sinai in 1844, Tischendorf also transcribed the folio containing the transition from Isaiah to Jeremiah, but he left this out of the 1846 publication. This transcription was only published after Tischendorf‘s second visit of 1853,4 during which visit Tischendorf was not able to gain access to the manuscript and found only a fragment of a folio of Genesis.5 It is not impossible that Tischendorf was simply refused access to the manuscript as, between his first and second visit, the manuscript was in all C. Tischendorf, Codex Friderico-Augustanus: sive fragmenta Veteris Testament e codice Graeco omnium qui in Europa supersunt facile antiquissimo in oriente detexit in patriam attulit ad modum codicis edidit C. Tischendorf (Leipzig: 1846). 2 Tischendorf made first mention of this number in 1865 in a “Mémoire” read to the Royal Society of Literature. See I. Ševčenko, “New Documents on Tischendorf and the Codex Sinaiticus”, Scriptorium 18 (1964): 55 and n. 2. 3 Folio 35.1 (OT 4) – 37.3 (OT 22), and folio 47.1 (OT 96) – 49.8 (OT 119). 4 C. Tischendorf, Fragmenta Sacra Palimpsesta, Monumenta sacra inedita. Nova collectio 1 (Leipzig: J.C. Hinrichs, 1855). 5 See e.g. Tischendorf‘s popular account of the discovery in the widely circulated little book When Were our Gospels Written? With a Narrative of the Discovery of the Sinaitic Manuscript (London: Religious Tract Society, 1896), 24. 1

5

6

History of Research

probability seen by the British Major MacDonald6 and certainly seen, and even studied, by the Russian archimandrite Porfiri Uspenski during his visits of 1845 and 1850.7 The latter discovered a loose fragment of a folio of Genesis8 and two fragments of a folio of Numbers, which were published by Tischendorf in 1865.9 Uspenski published part of the Genesis fragment and a part of the folio containing 1 Cor 13 in 1857.10 In this publication he revealed that the manuscript contained, besides the Greek Old and New Testament, also the letter of Barnabas in Greek. Apparently all the remains of the manuscript had been reunited between Tischendorf‘s first visit and the visits of Uspenski. In 1859, during Tischendorf‘s third visit to the Middle East, he obtained access to the manuscript and brought it to St Petersburg. A typeset semi-facsimile of the whole of the new discovery was published in 1862 in four parts; a less luxurious edition of the NT followed the next year.11 The Russian scholar Beneshevich recovered a new fragment of the codex from the binding of a book at Mt Sinai and published it in his

6 See the letter of June 20th, 1862, by Tregelles as published in T.C.F. Stunt, “Some Unpublished Letters of S.P. Tregelles Relating to the Codex Sinaiticus”, Evangelical Quarterly 48 (1976): 19. 7 According to Beneshevich, Uspenski published the account of his journey of 1845 in 1856, which included a long description of Sinaiticus. The following year, the account of his journey of 1850 appeared (see below n. 10). 8 The Genesis fragments found by Uspenski and Tischendorf come from the same folio. V.N. Beneshevich, Les Manuscrits Grecs du Mont Sinaï et le Monde Savant de l’Europe depuis le XVIIe Siècle jusqu’à 1927, (Athens: Verlag der “ByzantinischNeugriechischen Jahrbücher”, 1937), 64. 9 C. Tischendorf, Novum Testamentum Graece: Ex Sinaitico codice omnium antiquissimo Vaticana itemque Elzeviriana lectione notata edidit A.F.C. Tischendorf (Leipzig: 1865). 10 P. Porfiri, Vostok Khristianskii: Egipet i Sinai; bidy, ocherki, plany i nadpisi, 2 vols. (St Petersburg: 1857). The second volume contains the plates of which XV and XVI are a drawing of the passages of Sinaiticus. The 1 Corinthian 13 text covers 1 Cor 13:4–7, folio 82.4 (NT 73), line 3.18–32. 11 A.F.C. Tischendorf, Bibliorum Codex Sinaiticus Petropolitanus: Ex tenebris protraxit in Europam transtulit ad juvandas atque illustrandas sacras litteras edidit C. Tischendorf, 4 vols. (St Petersburg: 1862). Idem, Novum Testamentum Sinaiticum: sive Novum Testamentum cum epistula Barnabae et fragmentis Pastoris ex codice Sinaitico (Leipzig: F.A. Brockhaus, 1863). An extensive bibliography of Tischendorf is published by C. Böttrich, Bibliographie Konstantin von Tischendorf (1815-1874), (Leipzig: Leipziger Universitätsverlag, 1999).

Publication of the Codex

7

catalogue of manuscripts of St Catherine’s monastery.12 The fragment is a part of a folio of Judith, which was the only folio missing from an otherwise unbroken stretch of 116 completely preserved folios.13 A photographic facsimile of all the known text of Sinaiticus was published in 1911 and 1922 by Helen and Kirsopp Lake.14 In 1975, new parts of the codex were discovered in a previously blocked off room in the monastery. The 1999 catalogue of these finds contains photographs of two pages of Numbers. According to the description by Nikodopoulou, the new findings contain parts of Leviticus, Numbers, Deuteronomy, Judges, and Hermas. Besides a brief description of the contents, these findings have not yet been published. Tischendorf prefaced his 1862 edition with an introduction describing the physical and palaeographic attributes of the codex, which was taken over almost unchanged in the 1863 publication of the NT. In this introduction Tischendorf demonstrates the antiquity and importance of the codex but is especially interested in the text of the NT. Scrivener, in his “full collation” of Codex Sinaiticus of 1864, bases himself almost completely on Tischendorf and on the few pages published in facsimile.15 After Tischendorf, two other firsthand descriptions of the codex have been published: the introduction to the facsimile by Kirsopp Lake and, most important of all, the 1938 monograph by H.J.M. Milne and T.C. Skeat, Scribes and Correctors of the Codex Sinaiticus.16 In addition to the latter publication, the authors compiled a notebook in which all the corrections as listed in Tischendorf‘s edition were checked and, if needed, corrected. This notebook, in Skeat’s handwriting, is signed off by both authors on 12 V.N. Beneshevich, Catalogus Codicum Manuscriptorum Graecorum qui in Monasterio Sanctae Catherinae in Monte Sina Asservantur (St Petersburg: 1911 [reprint Hildesheim: Olms, 1965]). Beneshevich went on three occasions to Mt Sinai (1907, 1908, 1911) but does not tell when or from which book he recovered the Judith fragment. Beneshevich, Manuscrits Grecs, 93–94. 13 From the section folio 34.8 (OT 3) – folio 49.8 (OT 119) only folio 38.8 (OT 35) was missing. 14 H. Lake and K. Lake, Codex Sinaiticus Petropolitanus: The New Testament, the Epistle of Barnabas and the Shepherd of Hermas (Oxford: Clarendon Press, 1911). Idem, Codex Sinaiticus Petropolitanvs et Friderico-Augustanus Lipsiensis: The Old Testament (Oxford: Clarendon Press, 1922). 15 F.H.A. Scrivener, A Full Collation of the Codex Sinaiticus with the Received Text of the New Testament (Cambridge: 1864). 16 H.J.M. Milne and T.C. Skeat, Scribes and Correctors of the Codex Sinaiticus (London: British Museum, 1938). The study was made after the British Library acquired the manuscript in 1933.

8

History of Research

December 14, 1939 and is kept in the British Library under Add. Ms. 78935 A. The following table gives an overview of the modern locations of all the known parts of the codex: St Petersburg, Russian National Gen 23:19–24:19; Library; Greek codex 2 and 24:25–24:46 (fragments) Greek codex 259 Lev 20:27–22:30 Mt Sinai, St Catherine’s monastery; New collection, MG 1 Num 5:26–6:18; St Petersburg, Russian National 6:22–7:20 (fragments) Library; Greek codex 259 Num 16:7–20:28; 23:22–26:2 Deut 3–4 (fragments); 28:68–30:16; Judg 4:7–11:2

Mt Sinai, St Catherine’s monastery; New collection, MG 1

1 Chr 9:27–11:22

London, British Library; Add. Ms. 43725 Leipzig, Universitäts-Bibliothek; Greek codex 1

1 Chr 11:22–19:17; 2 Esd 9:9–end; Esth; Tob 1:1–2:2 Tob 2:2–end; Jdt 1:1–11:13

London, British Library; Add. Ms. 43725

Jdt 11:23–12:3; 12:5–9; 12:11–13; 12:14–16

St Petersburg, Russian National Library, Society of Ancient Literature; Ms. O. 156 London, British Library; Add. Ms. 43725

Jdt 13:9–end 1 Macc 4 Macc Isa Jer 1:1–10:25

Scribes and Correctors Jer 10:25–end Lam 1:1–2:20

9

Leipzig, Universitäts-Bibliothek; Greek codex 1

XII Prophets (Hos–Mal) London, British Library; Add. Pss–Job Ms. 43725 New Testament Barnabas Hermas, Visions 1–5; Commandments 1–4 (2, 3, 4 fragmentary) Hermas, Commandments 2–4 (fragmentary, St Petersburg, Russian National part of the inner column of a folio that is Library; Greek codex 843 for the most part in London) Hermas, Similitudes 6, 7:1–2, 8:14–16 and 18

Mt Sinai, St Catherine’s monastery; New collection, MG 1

SCRIBES AND CORRECTORS Tischendorf distinguishes four scribes working on the main text of Sinaiticus.17 Scribe A writes 1 Chronicles and 2 Esdras, 1 Maccabees and the last four folios of 4 Maccabees, the whole NT (except for six folios), and Barnabas. Scribe B writes the prophets and Hermas, scribe C the poetic books, and scribe D Tobit and Judith, the first half of 4 Maccabees, and in the NT, folios 74.2 (NT 10) and 74.7 (NT 15) in Matthew, 76.4 (NT 28) and 76.5 (NT 29) containing the ending of Mark and the opening of Luke, a folio in Thessalonians, 84.3 (NT 88), and in Hebrews, 84.6 (NT 91). Tischendorf does not know who writes most of the first column of Revelation but he notices a difference with the hand responsible for the rest of the book. A close similarity exists between scribes A and B, and between scribes C and D. Scribe D must be the main corrector of the text who carries out the diovrqwsiς. Folio 82.4 (NT 73) provides evidence that some of the corrections were made before the addition of the titles: the title in the upper margin is moved to the left because of an already existing correction. On the pages by scribe D in the NT there is only one correction by corrector A, which must be the scribe himself. A corrector is able to use more than one Tischendorf, Novum Testamentum Sinaiticum, xxi. Scrivener gives an English account of Tischendorf‘s description which is followed from here on. 17

10

History of Research

script, as is visible in the correction on folio 73.3 (NT 3), column 4. All the corrections are ascribed to one of the following correctors: A (contemporary, scribe D), B (described as a vir doctus), Ba (appearance somewhere in between A and B), and the series Ca, Cb, Cc, and Cc* (the last in Revelation only). Later and unimportant correction hands are D and E. Scrivener contrasts Tischendorf’s view of four different scribes with Tregelles’s opinion that “he could not find such diversity in the writing as would necessarily lead one to refer the several portions to different scribes.”18 Scrivener adds that, on the basis of the published facsimile plates, he would not endorse the view that multiple scribes had worked on the manuscript. However, the peculiar omega with its raised central stroke, which does not occur in the facsimile pages of scribes C and D, occurs eight times in the two pages ascribed to B, and less often in scribe A’s section. Additionally, Scrivener notes that the leaves of the NT ascribed to D are much freer from itacisms than those on either side of them. Scrivener, tentatively, suggests that perhaps we have to reckon with only two scribes: A and B, and C and D combined. Ultimately, the whole question of the number of scribes does not make much difference to Scrivener; the whole work was clearly executed at the same time, and transcribed from the same older copy.19 Scrivener accepts Tischendorf’s classification of the various corrector hands.20 In Sinaiticus the prima manus corrections are mostly recognizable by their in-line character (as opposed to later corrections, marked by erasure. The task of the diorqwthvς was fulfilled by scribe D, designated corrector Aa by Tischendorf but sometimes simply called corrector A. Scrivener expresses some doubt regarding the consistency of Tischendorf’s classification of individual corrections when he remarks that the corrector who fixed an omission in 1 Cor 13 used four examples of the peculiar omega, which did not really look like scribe D. Scrivener, relying on Tischendorf, describes corrector A as a hand as elegant as that of the prima manus, mainly correcting omitted words throughout the NT, except in Revelation. Only 15 corrections exist by a corrector called A oblique, all in the Pauline corpus. This hand leans slightly to the right and is somewhat inferior to corrector A, both in elegance and date. Corrector B is a full age below corrector A (end of the sixth century) and this corrector is followed by Ba, who only worked on the early pages of Matthew. The various C Scrivener, Full Collation, xvi. ibid., xviii. 20 ibid., xix–xxv. 18 19

Scribes and Correctors

11

correctors (C, Cb, Cc, Cc*) are all placed around the seventh century and again Scrivener notes that Tischendorf may not have been entirely consistent in assigning corrections to individual hands of the C group: “In the Gospels and Apocalypse Tischendorf indicates C by Ca, by way of distinction, reserving C for a few cases, wherein, we presume, he is doubtful by which member of this class the change was made.”21 In the introduction to the facsimile, Kirsopp Lake describes the various stages of work which occurred while the manuscript was still in the scriptorium:22 the writing of the text, the sub- and superscriptions, the tituli in Acts, the Eusebian canons, the stichoi numbering under the subscriptions to the letters of Paul, and the earliest corrections. Lake finds it difficult to distinguish between Tischendorf’s scribe A and B but accepts his judgement. The hand of scribe D is more characteristic. Lake sees the running titles in the NT as written by scribe D rather than scribe A, as Tischendorf thought, as they are also present on the folios by scribe D. Exceptions are the superscriptions in sections by scribe B, which are likely by the scribe himself. The actual form of scribe D’s superscriptions varies considerably. The subscriptions are written by the scribe of the main text, so that they are all by scribe A, except for Mark and 1 Thessalonians, which occur each on a folio written by scribe D. Also the subscription to John resembles A’s style much more than that of D, even though Tischendorf suspected that a different hand was at work in the last lines of the fourth Gospel. The hand which inserted the Eusebian apparatus is not assignable to any of the scriptorium hands and is therefore labelled by Lake as scribe E. The chapter divisions and tituli in Acts are all assigned to a corrector called A2. The stichoi in the Epistles are added in a small neat hand that is not identical with A2, but has more affinity with A4. Yet, Lake hesitates to equate these two as A4 looks “stiffer”; he therefore proposes a new scribe, “S”.23 Lake distinguishes more early correctors than Tischendorf. The A corrections are subdivided into the following types:24 A1 The uncial corrector A1 replaces the original writing with letters that are intended to resemble the erased ones as much as possible. As a ibid., xxiii. Lake and Lake, Codex Sinaiticus Petropolitanus: New Testament, xx–xxv. 23 ibid., xxii. 24 Unfortunately, the only place where Lake discusses the issue of multiple early correctors is in the Introduction of the facsimile. There he gives one page with examples of each hand (plate III), but never goes so far as to give a systematic overview of all corrections. 21 22

12

History of Research

rule, Tischendorf ascribes these to scribe D as sometimes the wedge sign is also added. But the inferiority of an erased surface makes it difficult to assign a hand with certainty. A2 Corrections showing alternation of small and large letters were ascribed to scribe D by Tischendorf who argued this on the basis of the correction in Matt 5:45, in which both uncials and small letters are used. Though Lake is not totally convinced by this particular example, he sees elsewhere in the manuscript reason enough to ascribe the A2 corrections to scribe D. A3 This corrector may or may not belong to the scriptorium. His corrections are marked by ligatures of h and n and an evenness in the size of the letters. A4 The few corrections labelled A4 are characterised by an extreme neatness and regularity and may be slightly later than the scribes mentioned so far. A5 This hand also made only a few corrections, which are marked by a very distinctive x. Aherm This hand seems to have acted as the diorqwthvς of Hermas only. The corrections known as A oblique, B, and Ba, differ from the hands above in that they probably do not belong to the scriptorium but cannot be later than the fifth century, which is earlier than the date assigned by Scrivener. A oblique An unimportant hand which made only a few casual alterations. B He made frequent corrections on the first pages of Matthew and specifically on orthography and is therefore very interesting. Ba The ink matches B, but the script A. It cannot be much later than the A group. Regarding the identity of the various C correctors, Lake notes that the ink of the author of extensive notes appended to Esther and Ezra shows the same reddish ink as that of the NT Ca corrector, but he does not believe the hands to be identical. Milne and Skeat looked afresh at the issue of the scribes and correctors of Sinaiticus and came up with a classification that partly supported Tischendorf but deviated considerably from the assessment by Lake.25 Their reinvestigation lead them to reduce Tischendorf’s four scribes to three, splitting up the poetic section originally attributed to scribe C 25

Milne-Skeat, 22–50.

Scribes and Correctors

13

among scribes D and A. The form of the kai-compendium (Â) and the occurrence of the filling mark, or diplè ( : ) are the main distinguishing hallmarks, apart from the general appearance of the script. The three scribes (A, B, and D) have each their own peculiarities. Scribe A starts to compress letters far back in the line, up to the sixth or seventh letter, and in poetic books even further back. Upright strokes in compressed letters are elongated, while rounded letters are very much diminished. He does not use the filling mark, and the oblique stroke of the kai-compendium makes an acute angle with the kappa. Scribe B rarely compresses letters further than three letters from the line ending. He makes a moderate use of the filling sign (though twelve times on folio 46.8 [OT 95]) and the angle of the kai-compendium is slightly more blunted. His script is also characterised by the far-right position of the superlinear stroke, which denotes a nu at line endings. His spelling is very poor. Scribe D uses compressed letters in much the same way as scribe A, but does not elongate the verticals. Also, he uses the filling mark more often than the other scribes and sometimes uses a double, or even triple version of it. The oblique stroke of the kai-compendium is rounded instead of acute. On the whole, his script is slightly smaller than that of the other two scribes and the size of the rounded letters as compared to the square letters is smaller. The first column of Revelation is assigned to this scribe. Milne and Skeat corroborate their case for the identification of these three hands by pointing out that each of the scribes uses a specific coronis at the end of a book. Though the coronis of a scribe may vary from book to book, it will always display a similar basic pattern. “The coronis, in fact, amounts to his signature, so distinctive is the design (or designs) adopted by each and so restricted the range of individual variation.”26 In subsequent research the identification of the three hands by Milne and Skeat has been widely accepted.27 Milne and Skeat reassign also the various tituli, superscriptions and subscriptions among the three scribes. The result is summarised in the following table. The numbers in the column “Running Titles” refer to the pages in Lake’s facsimile.28 Roman type indicates those that were probably written at the same time as the text, italic type indicates those probably Milne-Skeat, 27. So e.g. G. Cavallo, Ricerche sulla Maiuscola Biblica (Firenze: Le Monnier, 1967), 57 and B.M. Metzger, Manuscripts of the Greek Bible (New York: Oxford University Press, 1981), 76. 28 Taken from Milne-Skeat, 33–34. 26 27

14

History of Research

added later, and the + symbol indicates those located in the centre of the upper margin. book scribe superrunning titles subscripof scription tion text Gen (fragment) D .. B (?) .. Num (fragment) A .. .. .. 2 Esd A .. A (4, 6) A (including the section B (8, 10, 12, 14) of 1 Chr) Esth A A A A Tob D D B (24, 26) D D (28) Jdt D D D D 1 Macc A D .. A 4 Macc D D+ D (57 only) A A Isa B B+ B B Jer B B B B Lam B B .. .. Joel, Obad, Jonah, B B B Nah, Hab, Zeph Hag B B B (over Zechariah) B Zech B corrector (cf. Haggai) B Cb2 Mal B B .. B Pss D .. .. A A Prov A A B (176, 182, 184, A 186, 188) A (178, 180) Eccl A A B A Song A A A A Wis, Sir A A B A Job A A B A Matt A A+ A .. D

15

Scribes and Correctors book Mark Luke John Rom

scribe of text A D D A A A

superscription

running titles

subscription

D

A

D

D

A

A

A D+

A D

A A

In order to be able to present the evidence of the numbered epistles, the last three columns are split in two as the numbering of the epistles preceded the addition of the words. book scribe superrunning titles subscription script Wrd No Wrd No Wrd No 1 Cor A A A D (69, 72, 73) D A A A (70, 71) 2 Cor A D D D D A A Gal, Eph A D .. D .. A .. Phil, Col A A .. A .. A .. 1 Thess A A A .. .. D D D 2 Thess A A A .. .. A .. Heb A A .. A (omitting 91) .. A .. D 1 Tim A A A A (95), B (96) .. A A 2 Tim A A A .. .. A A Titus, A A .. .. .. A .. Phlm Acts A B+ .. B .. A .. Jas A .. .. .. .. A .. 1 Pet A A A A A A A 2 Pet A A A B (over 1 John) B A A 1 John A A A (cf. 2 Peter) .. A A 2 John, A A A .. .. A A 3 John Jude A A .. B .. A A Rev D D .. B .. A ..

16

History of Research

book

Barnabas Hermas

scribe A A B

superrunning titles script Wrd No Wrd

No

Wrd

No

A B+

.. ..

A ..

.. ..

.. ..

B B

subscription

The Eusebian apparatus was not assigned to any of the scribes by Tischendorf or Lake, but Milne and Skeat distinguish the work of two scribes:29 scribe A writes the first 52 numbers in Matthew, scribe D all the remaining ones, retracing also the numbers initially written by scribe A. The writer of the stichoi in the subscriptions to the Pauline letters uses the same small script as in the headings of the epistles and is therefore likely to have been scribe A. On the folio containing the end of 1 Thessalonians (copied by scribe D), no stichoi number is found, which is concordant with the fact that they were added by scribe A and not by the scribe of the cancel-leaf (in this case, scribe D). Besides the reassignment of the various scribal hands, Milne and Skeat also make a leap forward in the study of the various corrector hands. The A corrections are divided between corrections by scribes A and D,30 though Milne and Skeat are not always able to distinguish between the two. The larger corrections can be assigned on the basis of spelling, the shape of the arrow linking a correction in the upper or lower margin to the correct position in the text, the kai-compendium, and the form of the letter x. All the A oblique corrections are assigned to scribe D, including those corrections where a mixture of an upright and oblique script is used. Hallmarks of scribe A are the arrow with the rounded head attached to the shaft, the kai-compendium with the acute angle, and a fondness for ligatures. Scribe B only corrected his own work, which is not touched by any of the other scribes. The B and Ba corrections are also attributed to scribe A and D. The B corrections mainly occur on the first pages of Matthew and are to be seen as an abandoned project by scribe A to improve the quality of the codex.31 With the help of UV-light the problem of the last lines of the Gospel of John was solved.32 Initially the scribe finishes the gospel at the penultimate verse and places his coronis under the text. This coronis is afterwards erased, the final verse added and a new coronis Milne-Skeat, 36–37. Milne-Skeat, 40–44. 31 Milne-Skeat, 45. 32 Milne-Skeat, 12 and facing plate. 29 30

Scribes and Correctors

17

is inserted. All this is done by scribe A, who is also responsible for the remainder of the text. What Tischendorf thought was a different scribe turned out to be the same scribe writing over a large erasure, which does change the appearance of his script slightly. Milne and Skeat believe that most corrections are simply made against the exemplar from which the text was copied, though a few indications exist that at least some corrections, present in the exemplar, have found their way into the text. In Isa 28:22, for example, the two variant readings oi desmoi oi ofqalmoi are both in the text. Also Isa 24:7, where the scribe started o ploutos twn as- (i.e. asebwn, reading of Q), but corrected the last word to apantwn, and Isa 29:11, where the text reads toutouto, corrected by Ca to tauta, by Cb2 to touto. The group of C correctors is increased by Milne and Skeat.33 The Ca corrector is identified along the same lines as Tischendorf had done, and Cpamph is also recognised in agreement with Lake. With regard to the Cb corrections, Milne and Skeat base their case primarily on the OT, as the Cb corrections in the NT are all by one hand. Milne and Skeat distinguish Cb1, Cb2, and Cb3, of which Cb2 is the one also active in the NT. Cb1 corrects 2 Esdras, Esther, Judith, 1 and 4 Maccabees, and a couple of places in the Prophets. Cb2 is found in the Gospels and in Genesis, Numbers, and the Prophets (except Jer and Lam). Cb3 is found in the Prophets only and seems to make no positive contribution to the text; he simply removes additions and substitutions by previous correctors and touches the spelling of the original text. Cc’s hand is rougher than the other C correctors and uses a blacker ink. He is responsible for the section numbering in Isaiah and makes a single correction, but mainly rewrites or retouches text by the original scribe or by Ca. In the NT he only works on the first two pages of Revelation and Barnabas. His supposed correction of John 13:26 rests on a misprint in Tischendorf. Cc* is only found in Revelation where he finishes the task of Cc. In a review of Scribes and Correctors,34 Lake admits that he failed to note the change of hand between the first and second half of the Psalms but he has no doubt that Milne and Skeat are correct. However, he has still some doubts about whether the existence of a scribe C can be rejected. Concerning the identification of the scriptorium corrections with scribes A Milne-Skeat, 46–50. K. Lake, “Review of ‘Scribes and Correctors of the Codex Sinaiticus’”, Classical Philology 37 (1942): 91–96. 33 34

18

History of Research

and D, Lake also has reservations. Not all corrections in the A group are in the same script and though it is possible that scribes could use different scripts, it is, to Lake, not the most likely option.35

DATE AND PROVENANCE Tischendorf dates the manuscript firmly in the fourth century on the basis of the titles of the Gospels (a simple kata ... instead of a more extended title as euaggelion kata ...), the appearance of the script, and the contents of the NT, which includes Barnabas and Hermas.36 He is followed in this dating by the majority of subsequent scholars who give an opinion on this matter.37 A point of discussion remains, however, as to whether Sinaiticus could be part of Constantine’s order of 50 Bibles placed with Eusebius. If so, this would lead to a date in the second quarter of the fourth century and necessarily include Caesarea as the place of origin. The manuscript has also been dated in the latter half of the century, and in that case, we would not have a known context in which the manuscript was made.38

To a certain extent a similar disagreement is still reflected in NA27, p. 5, where the ℵ1 corrections are dated ‘4. - 6. Jh.’ This group of corrections presumably covers the A, B, and Ba corrections and is dated till up to two centuries after completion of the manuscript, while Milne and Skeat assign all these corrections to the original scribes. 36 Tischendorf, Novum Testamentum Sinaiticum, xxix–xxxiii. 37 Dean Burgon sought to argue that Vaticanus was at least a full generation earlier than Sinaiticus, but was succinctly refuted by Ezra Abbot, who showed that most of the arguments used for an earlier dating of Vaticanus were false. J.W. Burgon, The Last Twelve Verses of the Gospel According to S. Mark Vindicated against Recent Critical Objectors and Established (Oxford; London: James Parker and Co, 1871), 291–94. Ezra Abbot, “On the Comparative Antiquity of the Sinaitic and Vatican Manuscripts”, Journal of the American Oriental Society 10 (1880): 189–200. Gardthausen has a lengthy treatment on the date of Sinaiticus and prefers a date of 400 or possibly a decade or so later. His only palaeographical argument is the resemblance of the script with a dated inscription of AD 411. V.E. Gardthausen, Griechische Palaeographie (Leipzig: Veit & Comp., 1911–13), II: 122–34. Thompson (An Introduction to Greek and Latin Palaeography [Oxford: Clarendon Press, 1912], 200) dates Sinaiticus to the latter part of the fourth century, somewhat younger than Vaticanus. 38 On Constantine’s order of 50 Bibles see Eusebius, de vita Constantini, IV, 36– 37. Athanasius tells in passing that he himself sent copies of the Holy Scriptures to Constans (AD 341) but he does not inform us about the quantity of the order (Athanasius, Apologia ad Constantium imperatorem, 4). 35

Date and Provenance

19

In Scribes and Correctors, Milne and Skeat put forward an “almost incontrovertible” argument that the manuscript was written in Caesarea.39 In Matt 13:54 scribe A has the unique reading eis thn antipatrida. Rendel Harris had already pointed out in 1893 that this may well indicate local knowledge of the scribe, as the city Antipatris is only 45 km away from Caesarea.40 Milne and Skeat add that the same process may have been at work in replacing samareias with the incorrect kaisareias in Acts 8:5. As the first mention of Caesarea is only in Acts 8:40, this cannot be an instance of a reminiscence of an earlier mention of Caesarea. The form ippon for ioppon replaces the name of a town close to Caesarea with the name of a “well-known town on the south-west shore of the Lake of Galilee,” thus reflecting extensive local knowledge.41 The manuscript is dated by Milne and Skeat on the basis of a sequence created by eight examples of the specific bookhand used in Sinaiticus, called “biblical Uncial,”42 which suggest “that the Sinaiticus was written before the middle of the century”.43 Interestingly, Milne and Skeat add that if one had not known that all three scribes were contemporary, scribe D might well have been judged to be half a century later than scribe A and B. Kirsopp Lake holds a different opinion on the origin of the manuscript. In the introduction to the facsimile he argues for an Egyptian origin of the codex.44 He points out that each of the letters in Sinaiticus has its counterpart in the Egyptian papyri, but this evidence is inconclusive, “so long as we have no evidence as to other local hands.” The spelling krabaktos for krabattos and istrahleiths are both seen in Egyptian papyri. The link of Sinaiticus with Vaticanus is adduced by Lake as another strong link to Egypt, as the latter is shown by Rahlfs to have come from Alexandria.45 Additionally, the Euthalian sections in Acts reflect

Milne-Skeat, 66–69. J. Rendel Harris, Stichometry (London: C.J. Clay and Sons, 1893), 75. 41 Milne-Skeat, 68. 42 On the correct use of the terms ‘uncial’ and ‘majuscule’ see D.C. Parker, “The Majuscule Manuscripts of the New Testament”, in The Text of the New Testament in Contemporary Research, ed. B.D. Ehrman and M.W. Holmes, (Grand Rapids, Mi.: Eerdmans, 1995), 22. Uncial describes a particular Latin script, while majuscule is the term for a script ‘of a fair size’ as opposed to minuscule. 43 Milne-Skeat, 61. 44 Lake and Lake, Codex Sinaiticus Petropolitanus: New Testament, x–xv. 45 A. Rahlfs, “Alter und Heimat der vaticanischen Bibelhandschrift”, Nachrichten von der Königl. Gesellschaft der Wissenschaften zu Göttingen. Philologisch39 40

20

History of Research

the earliest form of Euthalius’s work, which is dedicated to Athanasius of Alexandria. Sanders disputes the early dating of the manuscript by Skeat, and proposes a date “nearer to the fifth century type [of manuscripts] than to the third.”46 The main argument is that as each scribe displays two or three different types of writing, the body of the text is likely to have been written in an archaistic hand as is described by Isidore (of Seville), Origines 6.14.47 To this should be added the running titles at the tops of pages and the extrusion of the first letter of a paragraph, both of which point to a later rather than an earlier date. Cavallo suggests in his study of the script that is used in Sinaiticus a date around 360, a decade later than Vaticanus.48 The palaeographical method Cavallo uses is based on the premise that the biblical Majuscule developed towards a perfect shape that was reached around the time of Sinaiticus and since then slowly degenerated. If one plots the datable examples one can deduce the non-dated specimens of the script. In 1999, Skeat returns to the issue and develops a fuller narrative of the origin of Sinaiticus.49 The similarity between the coronis of scribe D of Sinaiticus and scribe A of Vaticanus establishes the point that these two manuscripts come from the same scriptorium, as does the likeness in script and spelling.50 The case for Caesarea is fourfold: 1) the colophon of Esther in Sinaiticus, “which seems to show that in the seventh century at least Sinaiticus was at Caesarea.” 2) The curious reading, antipatrida,, of Matt 13:54. 3) The identity of hands in part of Sinaiticus and Vaticanus. 4) The curious chapter division in Acts, which can be traced through Euthalius to Pamphilus and Caesarea. Over against these four arguments, historische Klasse (1899): 72–79. Although Lake accepts the argument of Rahlfs, many scholars do not. 46 H.A. Sanders, “Review of ‘Scribes and Correctors of Codex Sinaiticus, H.J.M. Milne and T.C. Skeat’”, American Journal of Philology 60 (1939): 490. 47 Though Isidore writes around 600 AD, Sanders adds that it is an ‘obvious inference that the same condition prevailed in the earlier period’. ibid., 487. 48 G. Cavallo, Ricerche, 56–60. For an English discussion of Cavallo‘s argument see T.C. Skeat, “The ‘Codex Sinaiticus’, the ‘Codex Vaticanus’ and Constantine”, Journal of Theological Studies 50 (1999): 595–97. 49 Skeat, “The ‘Codex Sinaiticus’”, 583–625. 50 In Appendix C of Skeat’s collected Biblical writings, J.K. Elliott informs the reader that the similarity of the two hands ‘is not a point Skeat himself would now wish to dwell upon’. See The Collected Biblical Writings of T.C. Skeat (Leiden, Boston: Brill, 2004), 287.

Dictation

21

stands the observation by Rahlfs that the order of the books in Vaticanus resembles closely that in the Festal Letter of Athanasius of AD 367. Skeat’s discussion of Lake’s arguments in favour of an Egyptian origin of the manuscript lays bare the weakness of the palaeographical argument,51 but does not, of course, vindicate Skeat’s own position. According to Skeat, Sinaiticus must have belonged to the very first Bibles produced for Constantine52 but was rejected because of the uneconomical page format and the Eusebian apparatus. The size of a page of Vaticanus, which also belongs to the 50 Bibles for Constantine but was not rejected, allows for a much more economical use of the skin of a goat, allowing a saving of almost 50% in the number of skins needed. The Eusebian apparatus was not only left out because the canon tables would take up considerable space but also because Eusebius, and therefore his newly devised system, would want to avoid criticism from the ecclesiastical authorities or the emperor.53 Work on Sinaiticus was broken off just before it was completed and, according to Skeat, it lay on a shelf in the scriptorium for two centuries till it was corrected in the days of the C correctors.

DICTATION In the field of codicology and the study of book production, Sinaiticus is known for its connection with the dictation theory. In Scribes and Correctors, a whole chapter is dedicated to a study of the orthography of the manuscript.54 On the basis of this study, Milne and Skeat propose that Sinaiticus has actually been produced by dictation, that is, a reader read the text aloud to a number of scribes who each simultaneously copied the text into their own codex. Before Milne and Skeat, discussion on the role of dictation in manuscript production was already going on,55 but Sinaiticus was the first biblical manuscript for which dictation was actually argued on the basis of specific data.56

Skeat, “The ‘Codex Sinaiticus’”, 586–92. ibid., 612–17. 53 ibid., 615. 54 Milne-Skeat, 51–59. 55 See the overview of the discussion in T.C. Skeat, The Use of Dictation in Ancient Book Production, Proceedings of the British Academy 1956 (1957), 179–90. 56 Petitmengin and Flusin have argued that dictation was used to produce a Syro-Hexapla. They base their case on a letter of the patriarch Timothy I, written in 796/7, which mentions two readers and six scribes who were to produce three manuscripts. P. Petitmengin and B. Flusin, “Le Livre Antique et la Dictée: 51 52

22

History of Research

The main argument for the dictation theory is the orthographic fluidity of Sinaiticus. The differences in spelling between the scribes cannot be explained by means of different exemplars, and scribe B produces such a badly spelled text that it is hardly conceivable that this could have been possible if he had a correctly written text in front of him. A few specific instances from the text are discussed in support of the dictation theory. The most important example is from 1 Macc 5:20 where the text reads h_§h_g (the gamma looks in the manuscript more like ‚i- ), which is interpreted by Milne and Skeat as h ex h triscilioi.57 The normal text reads oktakiscilioi (which would be just h_), and the current text can best be explained by the fact that the reader called out “either six or three thousand” which was subsequently copied by the scribe into the text. If this variant already existed in the exemplar, the only consequence would be that that exemplar must have been dictated. Other examples may be seen in passages which have been duplicated because the scribe made a leap back, such as Jer 4:7 where kaqereqhsonte appears the first time and katereqhsontai the second (scribe B), and 1 Thess 2:13 where we have energeitai, umis, mimhtai the first time and energitai, umeis, mimhqhte the second time (scribe A).s An objection to the dictation theory is that it would invite an enormous amount of phonetic errors to be made. Milne and Skeat answer that this risk was indeed taken, sometimes with disastrous results (as in scribe B). Scribe D, however, provides evidence that a well-trained scribe could produce a text with only a very small percentage of error. As a result, when weighing variants which are phonetically identical, one cannot claim passages written by scribe A or B as witness for or against a particular reading. Both Sanders and Lake have aired their doubts regarding the dictation theory.58 However, Skeat reiterates his view in later publications,

Nouvelles recherches”, in Mémorial André-Jean Festugière: Antiquité paienne et chrétienne, ed. E. Lucchesi and H.D. Saffrey (Genève: 1984), 247–62. 57 1 Macc 5:20, folio 39.8 (OT 43), line 1.19. Milne-Skeat, 57. 58 Lake, Review: 95. ‘(I) am inclined to hold that the argument from orthography is nearly worthless. Some people can copy correctly, others cannot.’ Sanders, Review, 487–88: ‘But if that [viz. phonetic errors in spelling—DJ] constitutes proof, then all the manuscripts that I have ever studied, were written from dictation.’

Dictation

23

though adding that perhaps not enough attention was given to the possibility of “subconscious dictation.”59 The French scholar Alphonse Dain gives, in his introduction to codicology, an analysis of the actual copying process, which has become very influential in NT textual criticism.60 Copying is broken down into four separate actions: reading of the exemplar, retention of the text, internal dictation, and finally the action of the hand (jeu de main). Dain is not convinced by Skeat’s orthographic argument: some form of dictation always plays a role in the copying process.61 Dain is followed by Klaus Junack, who addresses the issue of scribal habits and dictation following Dain’s model of copying.62 He emphasises that in the scriptio continua of classical Greek manuscripts an acoustic rendering of the text is necessary to extract its meaning. As words are not separated from one another, a reader or scribe has to derive the meaning from the sound of the syllables. Because the phonetic character of the text is so important, it is not strange that some scribes copy with more orthographic errors than others.63 59 Skeat, Use of Dictation, especially 191–93. Skeat himself may perhaps show some doubt when he calls the whole discussion on dictation ‘difficult’. Skeat, “The ‘Codex Sinaiticus’“, 584 n. 6 ‘I do not want to discuss here the difficult question of dictation’, possibly because of Klaus Junack‘s work (see below). Milne-Skeat are followed by Tom Pattie, “The Codex Sinaiticus”, The British Library Journal 3 (1977), 5. 60 A. Dain, Les Manuscrits (Paris: Les Belles Lettres, 1975), 41–45. This analysis of the copying process is also used by B.M. Metzger, The Text of the New Testament, 3rd ed. (New York; Oxford: Oxford University Press, 1992 [first ed. 1963), 16 (without reference to Dain). 61 Dain, Les Manuscrits, 21. 62 K. Junack, “Abschreibpraktiken und Schreibergewohnheiten in ihrer Auswirkung auf die Textüberlieferung”, in New Testament Textual Criticism, ed. E.J. Epp and G.D. Fee (Oxford: Clarendon Press, 1981), 277–95. 63 Modern psychological research into reading assumes that in all reading phonological aspects play some role as the mental lexicon seems to be organised both on sound and word image (morphophonemic). See e.g. R.G. Crowder and R.K. Wagner, The Psychology of Reading (New York; Oxford: Oxford University Press, 1992), 157–88 and G. Underwood and V. Batt, Reading and Understanding (Oxford: Blackwell, 1996), 12–21. This makes Junack‘s reliance upon Balogh unnecessary (J. Balogh, “Voces Paginarum”, Philologus 82 (1927): 84–109, 202–40). Balogh argued that all reading in antiquity was loud reading and that silent reading was a high exception. His arguments were questioned by B.M.W. Knox, “Silent Reading in Antiquity”, Greek, Roman, and Byzantine Studies 9 (1968): 421–35. Junack may well be right, though, that internal pronunciation becomes even more important when the word image is not immediately recognisable such as is the case with scriptio continua,

24

History of Research

Most recently, Chrys Caragounis accepts that copying by dictation must have been a breeding place for unconscious and involuntary orthographic errors.64 However, Caragounis admits that similar errors can arise in situations where the scribe works all on his own. The main implication that can be drawn from such spelling diversions is that such scribes are anorthographoi (= incapable of spelling correctly) and lacking solid education.

TEXT Most of the studies on the textual character of Sinaiticus fall outside the scope of our study as they have been carried out to establish the value of Sinaiticus for the textual criticism of a particular book and do not shed any light on the manuscript or its scribes. The remarks which Tischendorf and Westcott-Hort have made on the singular readings of Sinaiticus will be discussed in Chapter 5. Here we will mention briefly some studies that possibly shed some light on the scribal behaviour. In a short note, H.N. Bate suggested that in Luke 22:40 (genovmeno" deV ejpiV tou’ tovpou ei^pen aujtoiv", proseuvcesqe mhV eijselqeivn eij" peirasmovn) the reading of Sinaiticus, the infinitive proseucesqai for the imperative proseucesqe, could well be the original reading.65 Bate had noticed this reading of Sinaiticus because the manuscript lay open at this verse in the British Museum. He argued that not only is the infinitive good Lucan style, but the reading has also the support of W Q N 13 506 and D (with some variation in the following text), and should be recorded in the critical apparatus, something that von Soden and Tischendorf fail to do. The stylistic difference with verse 46 may very well have been intentional. Milne and Skeat refer to this note by Bate as an example of where the “fact that the Sinaiticus was undoubtedly written from dictation” should be taken into account when assessing phonetically identical readings.66 The discussion is interesting because it raises the issues of dictation and the phonetic interchange between ai and e. We will come back to the latter issue in Chapter 3. see H.Y. Gamble, Books and Readers in the Early Church (New Haven; London: Yale University Press, 1995), 203–5. 64 C.C. Caragounis, The Development of Greek and the New Testament (Tübingen: Mohr Siebeck, 2004), 494–502. 65 H.N. Bate, “Luke xxii 40”, Journal of Theological Studies 36 (1935): 76–77. Luke 22:40, folio 78.5 (NT 45), line 7.25. 66 Milne-Skeat, 59 and n.1.

Text

25

In 1932, Alfred Rahlfs published a short note in which he pointed out two hexaplaric marginal notes that had crept into the main text.67 The first is found in 2 Esd 13:15, where Sinaiticus reads kolumbhqras twn qetou silwam,68 which should be interpreted as an intruded marginal note giving the reading of Theodotion and the Quinta (q'e') for kolumbhqras which is tou silwam. A similar phenomenon occurred in 2 Esd 22:27 where Sinaiticus reads en qwlaqas . en exomolghsei69 for en qwlaqa. The intruded note must have read s' en exomologhsei, which introduces an alternative reading of Symmachus. Though the 2 Esdras text itself does not show other influences of the Hexapla, apparently marginal notes were present in the exemplar of Sinaiticus or in one of its ancestors. Gordon Fee studied the first nine chapters of the Gospel of John70 and found that Sinaiticus is a leading witness to the Western text in the first eight chapters (up to John 8:38). Though Fee gives no explanation as to how this could have happened, it seems that some sort of block mixture has gone on. A posthumously published study by Christian Tindall attempts to merge literary source criticism and statistical analysis of a manuscript.71 After counting the number of letters in each column, large divergences from the average are noted and related to the actual text of these deviating columns. By this method twenty additions to the original text of the Gospels are identified. The tacit presupposition of Tindall’s method is that the scribe of Sinaiticus had access to a version of the Gospels which was more pure than that from which any of the other manuscripts derives. It is not remarkable, therefore, that this study has been by and large neglected, save for the appendix in which all the letter counts are presented. However, A. Rahlfs, “Curiosa im Codex Sinaiticus”, Zeitschrift für die alttestamentliche Wissenschaft 50 (1932): 309–10. See also R. Hanhart, Text und Textgeschichte des 2. Esrabuches (Göttingen: Vandenhoeck & Ruprecht, 2003), 299–300. 68 2 Esd 13:15, folio 35.7 (OT 10), line 1.15. 69 2 Esd 22:27, folio 36.4 (OT 15), line 3.40. Rahlfs prints a raised dot after qwlaqa”, which is present in a supralinear position but is not necessarily original. 70 G.D. Fee, “Codex Sinaiticus in the Gospel of John: A Contribution to Methodology in Establishing Textual Relationships”, in Studies in the Theory and Method of New Testament Textual Criticism, ed. E.J. Epp and G.D. Fee, (Grand Rapids, Mi.: Eerdmans, 1993), 221-43, originally published in New Testament Studies 15 (1968/9), 23–44. 71 C. Tindall, Contributions to the Statistical Study of the Codex Sinaiticus (Edinburgh: Oliver and Boyd, 1961) 67

26

History of Research

A.Q. Morton, who had helped in the publication of Tindall’s work, revisits the approach in 2002.72 Though Morton tries to refine some of the details of the statistical work of Tindall, the basic assumption remains the same: editorial work on the text of the Gospels can be identified by deviations from the average number of letters per column. Both Morton and Tindall seem unaware of the codicological description by Milne and Skeat. Though the assumptions of Tindall’s study are easily countered, there is a sense in which some of the studies in Chapters 2 and 3 below are related to the type of question Tindall raises. There we will also look at the actual formatting of the text in order to deduce things that were going on behind the scenes. However, we will not be looking for a large-scale redaction of the biblical text, but rather for clues as to how the scribes interact and produce a large codex.

HISTORY OF RESEARCH AND THIS STUDY In this present study, we will not repeat the work done by Milne and Skeat, but take their investigation one step further and investigate the role and peculiarities of the three scribes. Milne and Skeat studied the manuscript in order to learn about Sinaiticus; we will study the manuscript in order to learn about the scribes, their habits, and the influence of the individual scribe on the final form of the text. Then we will ask what we can learn from this about the scribal behaviour in general? Besides this, the issues of dating and provenance are not solved satisfactorily for Sinaiticus, and even though these questions are not our main interest, we will examine them in the final chapter, in which we evaluate the results of our study against the Caesarean hypothesis of Skeat. Some of the non-textual features of the codex may assist us in determining more precisely the setting in which the codex was produced. Also, the prominent question of the dictation theory still remains without solution. Can Sinaiticus really be put forward as a manuscript written by dictation? Several of the individual studies in the next chapters provide fresh data which, upon reflection, will provide sufficient arguments to settle this question. In Chapter 2 we will reflect on the codicological composition of Sinaiticus, the interaction of the scribes and the distribution of the various scribal tasks among them. Chapter 3 contains a series of studies on different scribal practices in the text. We will look at the nomina sacra, ligatures, spelling, the Eusebian apparatus, and paragraph division. These studies will 72

14–31.

A.Q. Morton, “Codex Sinaiticus Revisited”, Irish Biblical Studies 24 (2002):

History of Research and This Study

27

serve both to learn more about the individual scribes and to determine the level of control exercised by the environment in which Sinaiticus was written. The important question of scribal influence on the wording of the text will be discussed in Chapter 4. We will study scribe D and scribe A in order to establish their scribal habits as regards the text: Can we say anything about the quality of the copying process? And to what extent are the scribes responsible for introducing readings into the text that were not present in their exemplar?73 The final chapter serves to reflect upon the collective results of the individual studies and tries to place these in the context of past discussions of Sinaiticus.

These are the true copying errors, which is a different question from whether a reading is ‘original’ or not. 73

2 THE NON-TEXTUAL CHARACTERISTICS OF SCRIBAL BEHAVIOUR IN CODEX SINAITICUS

INTRODUCTION Any written text is formatted to fit the particular medium on which it is written. A papyrus roll demands a different approach than an ostracon, and likewise a codex consisting of multiple vellum quires requires decisions on the size of the letters, number of columns per page, the width of the margins, and the number of folios per quire.1 In Sinaiticus most of these elements are very regular, showing only the occasional variation. Lines do vary in length, columns contain normally 48 lines but sometimes fewer or more, the number of columns per page is mostly four or two, and a quire has normally eight folios as it is made out of four folded sheets. The question is whether such variation is simply random and a result of the whim of the scribe or whether it might reveal something of what is going on behind the stage curtains, as it were, of codex production. Some of the formatting is solely the responsibility of the scribe or of the designer of the codex, who may be a patron or purchaser of the codex: nothing inherent in the text requires it to be written in a certain number of lines per column. Only the length of the text that is copied may be of influence. However, some elements in the formatting are strongly suggested by the text or its transmission history. One could think for example of the major breaks that show the start or end of a book or group of books, such as the poetic books of the OT or the Pauline corpus; the marking out of each individual psalm in the Book of Psalms by means of a title and number; the singling out of certain structural elements within a book, such as some of the visions in Isaiah, which receive a distinct title; or the use of lists for the genealogies in the gospels. Possibly paragraphing could have been added to this list. We will deal with paragraphs in Chapter 3. 1

29

30

Non-Textual Characteristics

The scribes are only accessible to us through their work in the codex. It is for this reason that we start with the study of the physical elements that constitute the manuscript. In this chapter the interplay between these elements, such as page, sheet, and quire, and the scribes who wrote the text will be considered more closely.

THE CONSTITUTIVE ELEMENTS Sheet and quire Sinaiticus is one of the earliest remaining examples of an extensive multiquire vellum codex,2 though the technique of adding multiple quires upon one another does not seem to postdate the creation of single-quire codices by much.3 A quire is formed by placing a number of sheets upon one another and then folding the whole stack together. The formation of a single quire is described in Scribes and Correctors, though the information provided there cannot be checked any more since the rebinding of the manuscript in the 1930s.4 The rule followed in Sinaiticus is that four sheets of parchment (a quaternion) are used and that the hair side of one sheet of parchment always faces the hair side of a previous sheet and likewise the flesh side always faces the flesh side of another sheet, the result being that wherever one opens the codex facing pages showed the same side of the 2 The earliest multi-quire papyrus codex I am aware of is P66 (P. Bodmer II ca. AD 200), which, according to Kasser, who revised the reconstruction of the quires as made by V. Martin, the original editor, must have contained 8 quires of irregular length to contain the whole Gospel of John, see E.G. Turner, The Typology of the Early Codex, (University of Pennsylvania Press, 1977), 70 n. 12. The fragment of P66 in Cologne shows the central opening of the seventh quire and may well confirm the reconstruction of the final three quires of P66. Other early examples are P45 (P.Chester Beatty I) from the third century, and P.Chester Beatty II (dated second/third century by C.H. Turner). The earliest multi-quire parchment codex is P.Ryl. 1 53, dated third/fourth century, a Homer codex with a quire number still visible. 3 See the discussion in Turner, Typology, 55–71. Kurt and Barbara Aland seem to regard the development of adding multiple quires as a secondary development from as late as the early third century; K. Aland and B. Aland, The Text of the New Testament, 2nd ed. (Grand Rapids, MI; Leiden: W.B. Eerdmans; E.J. Brill, 1989), 75. 4 Milne-Skeat, 72–73. For an overview of the types of quires see besides Turner, Typology, also J. Irigoin, “Les cahiers des manuscrits grecs”, in Recherches de Codicologie Comparée, ed. Philippe Hoffmann (Paris: Pr. de l’École normale supérieure, 1998), 1–19.

The Constitutive Elements

31

parchment.5 The aesthetic reasons for this are obvious, as the flesh side of the parchment is initially smoother and whiter. The folded quire always has the flesh side at the outside.6 Regular Quire

f-h . h-f . f-h . h-f | f-h . h-f . f-h . h-f

In Sinaiticus, a few times a book starts on a new quire with the previous quire being shortened. Two folios must have been present in the last quire of Malachi, the last book of a long section written by scribe B, one folio at the end of Luke, and two folios at the end of John.7 All these superfluous folios have been cut out near to, but not along, the actual folding of the sheet so that just little stubs remain.8 In Vaticanus the same phenomenon occurs only in the transition from Tobit to Hosea, which is the transition from the Wisdom books to the Prophetic books. There the last page of quire 49 has been cut out. A different scribe continues with Hosea on quire 50.9 Of the other big biblical codices Alexandrinus is also formed by quaternions, while Vaticanus consists of quinions, having five sheets per quire. 6 The combination of the flesh side on the outside of a quire and like facing like is commonly known as ‘Gregory’s Law’. This rule is generally followed in parchment manuscripts from the East. The Latin west seems to have followed a different pattern: hair always on the outside of a quire (C.R. Gregory, “The Quires in Greek Manuscripts”, American Journal of Philology 7 [1886]: 27–32). The early papyrus codices did not always follow the rule of like facing like (Turner, Typology, 65). 7 Milne-Skeat, 100, 107, 108. 8 This is already noted by Gregory in “Quires in Greek Manuscripts”. Jean Irigoin (“Les cahiers des manuscrits grecs”, 13) observes that in medieval manuscripts the quire size was often adjusted to make sure that the end of a work would fall within the limits of a quire so that the start of a new work would coincide with the beginning of a quire. For a similar practice in Sinaiticus see below p. 43. McGurk notes that often in Greek and Latin Gospel books each Gospel is given its own separate set of quires. P. McGurk, Latin Gospel Books from A.D. 400 to A.D. 800 (Paris: Editions “Erasme” Standaard-Boekhandel, 1961), 8–9. 9 In Vaticanus the three books Esther, Judith, and Tobit follow Sirach and are probably seen as belonging to the Wisdom books. See also P. Canart, “Notice Paléographique et codicologique”, in Prolegomena to Bibliorum Sacrorum Graecorum Codex Vaticanus B (Roma: Istituto Poligrafico e Zecca dello Stato, 1999), 3. 5

32

Non-Textual Characteristics

The physical central fold of the quire does not change by cutting out one or two leaves at the end. When quires are bound together into one volume, they must be sown through the central fold of each quire. A feature of Sinaiticus which has escaped notice so far is a set of markings in the outer margin of both pages of many of the central openings of the quires. It looks a little like a zeta on top of a xi. These marks are also present on the physical central opening of those quires that are shorter because the last leaves have been cut out, as well as on the central opening of the three quires that are shorter because fewer sheets were used to compose the quire.10 These markings can only have served the binders of the manuscript. However, given the heavy hue of the ink it is unlikely that these markings come from the original scriptorium. It is more likely that they belong to the rebinding of the manuscript, possibly connected with the so-called C correctors who worked sometime between the fifth and eighth centuries.11 The quires have two numbering schemes. In many quires, a number is visible on the top left corner of the first page.12 Milne and Skeat take this numbering to be the original numbering and assume that in many places it has been cut off during a rebinding of the manuscript. The second numbering, dated to roughly the eighth century,13 is placed in the top right corner of the first page of a quire and is visible on each quire. In the OT, the quire numbers of the two numberings are always in agreement with one another, but in the NT the older quire numbering is consistently one number higher than the more recent one. Milne and Skeat explain this The presence of these markings could have prevented the mistaken conception that six leaves were missing from quire 91, consisting of only one folded sheet (see e.g. F.H.A. Scrivener, A Plain Introduction to the Criticism of the New Testament (London: G. Bell, 1894), 97 and M. Hengel and R. Deines, The Septuagint as Christian Scripture (Edinburgh; New York: T & T Clark, 2002), 58. As the markings are present on the central opening it is clear that nothing has been cut out from the back of this quire. 11 Brigitte Mondrain mentions two examples of codices in which the centre of a quire is marked by either the repetition of the quire number or the word meson. Her examples date from the fourteenth and fifteenth century B. Mondrain, “Les signatures des cahiers dans les manuscrits grecs”, in Recherches de Codicologie Comparée, ed. Philippe Hoffmann (Paris: Pr. de l’École normale supérieure, 1998), 23–24. 12 Mondrain does not know of any manuscript in which the quire signature is placed in the upper-left corner and fails to mention Sinaiticus. She assumes erroneously that the signature in the upper-right corner is original (“Les signatures”, 25). 13 Milne-Skeat, 7. 10

The Constitutive Elements

33

discrepancy by assuming that a quire was planned between Job and Matthew, which was to contain the Eusebian canon tables, but was actually never fitted in.14 As the Eusebian apparatus is not complete in the Gospel of Luke, Milne and Skeat assume that the scribes did not insert the Eusebian tables at all: the project was simply abandoned. The suggestions that a simple mistake in the numbering was made or that a quire containing some OT apocryphon or the Eusebian canon tables had somehow been lost are rejected. Though Milne and Skeat certainly paint a plausible scenario, the argument relating to the unfinished character of the apparatus in Luke has only limited force as the apparatus was inserted in John, which follows Luke. Patrick McGurk notes that in many Eastern Gospels and in some Latin Gospel books, the quire containing the introductory material was not numbered.15 If this practice was observed during the renumbering of the quires, it is even possible that a quire with the Eusebian tables was still present between Job and Matthew at the time of the renumbering and was lost much later. Whatever the case, the question of the missing quire that possibly never was will remain undecided unless new information comes to light. The size of the individual sheet is very large compared with other surviving manuscripts, and must have originally measured around 16 (H) by 29 (W) inches (around 41 by 74 cm).16 There has been some debate over the type of animal that provided the skin for the parchment, especially since Tischendorf’s suggestion that antelope skin was used. Gregory doubts whether enough young antelopes were around in Palestine to provide for the necessary skins.17 Milne-Skeat, 7–9. P. McGurk, Latin Gospel Books, 8 and n. 3. 16 Milne-Skeat, 71. For the dimensions of other parchment codices see Turner, Typology, 26–30. Sinaiticus is the largest of the four early parchment codices with a single page measuring 41 by 37 cm. A page in Vaticanus measures 27 by 27 cm. Alexandrinus and Ephraemi Syri Rescriptus 32 by 26 cm and 33 by 27 cm. See K. Aland, Kurzgefasste Liste der Griechischen Handschriften des Neuen Testaments, 2nd ed., (Berlin; New York: Walter De Gruyter, 1994), 19. 17 C.R. Gregory, Canon and Text of the New Testament (Edinburgh: T. & T. Clark, 1907), 333–34. DNA analysis of some of the Dead Sea scrolls show that some of the leather roles were made from the skin of the ibex, which at least shows that this animal was used for the production of writing materials. D.W. Parry et al., “New Technological Advances: DNA, Databases, Imaging Radar”, in The Dead Sea Scrolls after Fifty Years, ed. P.W. Flint and J.C. VanderKam (Leiden: Brill, 1998), 506. 14 15

34

Non-Textual Characteristics

The quality of the parchment used for Sinaiticus is very high. Tom Pattie notes that very few leaves have holes in them, which means that the parchment must have been selected carefully.18 Still, some original holes in the parchment do occur every now and then; these can be recognised by the fact that the writing continues without interruption in the text at the other side of the disturbance. Milne and Skeat describe how a quire was pricked and ruled. By pricking tiny holes through the sheets, the anchor points for the actual ruling were created. The vertical lines for the width of the columns was set by the pricking in the upper and lower margin, while a series of pricks from top to bottom set the anchor points for the horizontal ruling. The quire was provisionally fastened together and laid open upside down, the outside sheet was ruled over both pages as a whole, and subsequently each opening showing the flesh side was ruled, thus ruling pages of two different sheets at the same time (except for the central opening of a quire). By doing this the result was that on each opening showing the flesh side both facing pages had exactly the same ruling. Care had to be taken that the sheets in the quire did not move during the pricking and ruling. In the first half of the codex this was achieved by whipping the quire over with crossing stitches of narrow vellum strips; for the latter half the technique was changed to gluing the outer three sheets onto the central sheet with thick glue. To do this, an incision of about 1–1½ inches was made into the fold at its head and tail.19 The change of technique halfway through shows not only that the procedure was not standardised from the outset, but also that each of the two halves of the manuscript was produced in order. The pattern of the horizontal ruling differs from quire to quire: sometimes each line is ruled, sometimes every other or every third line, and often we find combinations of these. It has not been possible to link these diverse patterns with individual scribes. Folio and page Each individual leaf has a hair and flesh side; the position of a leaf in a quire determines which of the two forms the recto. Despite the fact that the flesh side initially looks brighter, it suffers more over time and nowadays the T.S. Pattie, “The Creation of the Great Codices”, in The Bible as Book: The Manuscript Tradition, ed. J.L. Sharpe III and K. Van Kampen (London; New Castle: British Library; Oak Knoll Press, 1998), 64. 19 Milne-Skeat, 73. Exceptions are noted as well: the outer sheet of quire 40 is ‘foreign’, the cancel-leaves in the NT do not have the ruling of that of the rest of the quire, and quire 90 shows individual ruling of each sheet. 18

The Constitutive Elements

35

flesh pages are slightly more difficult to read than the hair pages. The hair side tends to absorb the ink better than the flesh side, which results in a much better preservation of the actual writing in the long term.20 Each page needs preparation before writing. The parchment needs to be pounced (i.e., the surface needs to be roughened up so that the ink will hold). Milne and Skeat show that this pouncing was not done immediately after the quire was prepared, but only just before the scribe was about to write on the actual page.21 This is shown by quires 42 and 62, where scribe A takes over from scribe D within a quire. In both instances the change in the pattern of pouncing coincides with the change of scribe. Scribe D is described as using a strictly perpendicular pattern, while scribe A has a typical swerve in the upper margin.22 In a number of the folios in the second half of quire 62 this pattern is indeed still visible, just as scribe D’s pattern is occasionally noticeable before the break. I was not able to distinguish the two different patterns on the folio of the transition from scribe D to scribe A.23 In a few instances the page was the unit that determined the break between books. Matthew ends on the first line of the final column of a recto page, but it would have been relatively simple to have ended the book on the previous column. It may well be that the last column of the page was used in order to let the Gospel of Mark commence on the next page, the verso, without leaving a column blank. Likewise, the book of Acts finishes on the third column of the recto, but here the last column is left blank, and the letter of James starts on the verso. In Vaticanus the rule is that the next book commences on the next available column,24 in Alexandrinus the transitions between the apostolicon, the Pauline corpus, Revelation, and 1 and 2 Clement are well indicated by blank columns or pages. In Alexandrinus,

Gregory, Canon, 324. Milne-Skeat, 79. 22 Milne-Skeat, 79. 23 In the conservation assessment carried out in the course of the Codex Sinaiticus Digitisation Project special attention will be given to the pouncing patterns. With the correct light fall the pattern of pouncing is reasonably well visible in the upper and lower margin but is not always detailed enough to establish individual patterns. 24 ‘Les pièces s’enchaînent dans le manuscrit d’une manière qu’on pourrait qualifier d’ “économique”‘. Canart, “Notice Paléographique et codicologique”, 3. The exceptions in Vaticanus are the transition from Tobit to Hosea and the open column before the gospel of Luke. 20 21

36

Non-Textual Characteristics

the Gospels are also separated from one another by one or more columns of chapter headings, the capitula.25 Not just the page but also the folio could serve as the unit which determined the demarcation between books or group of books. In Sinaiticus, the letters of Paul end with two blank columns on the recto and a blank page on the verso. The book of Acts commences on the next folio, which is the seventh of quire 85. Column and line Barring the influence of book endings that were given empty columns, the normal page has four columns in the case of prose books and two columns in the case of the poetic books. The poetic books are written in sense lines, in which each poetic verse line starts on a new line. A result of the twocolumn format is that a line in a column can often contain the whole stichos.26 Whenever a stichos does not fit on a single line it overflows to the next line(s), which is indented by four or five letters.27 The alignment of poetic lines to the right margin is much less strict than for prose. Therefore it often happens that a line is broken off after a full word, but this is certainly not always the case. Normally, care is taken that if an overflow line (or lines) is needed, a sufficient amount of text is put on such line; if necessary the number of letters on the previous line is reduced. The poetic format is also followed in Vaticanus where the normal layout for prose texts (three columns) is reduced to two columns for the poetic books. The following books are written in the poetic format: Psalms, Proverbs, Ecclesiastes, Song of Songs, Wisdom, Sirach, and Job. The prologue to Sirach, however, is treated as prose and is written in a column that is narrower than the normal poetic column, but broader than the indented poetic line. Again, this is in line with the way Vaticanus deals with the same text. In Alexandrinus two instances are found in which the next book within a collection continues in the same column where the previous ended (once in the Catholic epistles and once in the Pauline corpus). 26 For a discussion of the term stichos see D.C. Parker, Codex Bezae (Cambridge: CUP, 1992), 73–75. 27 P.Bodmer IX (early fourth century) shows the same layout in stichoi in one column per page, P.Bodmer XXIV (date disputed; second to fourth century, see A. Pietersma, “The Edited Text of P.Bodmer XXIV”, Bulletin of the American Society of Papyrologists 17 [1980]: 67–79) and P. Chester Beatty XIII and XIV (both fourth century) indicate the next stichos by a short space or a colon. Both systems, then, were current side by side in the fourth century. 25

The Constitutive Elements

37

Normally a column contains 48 lines, but exceptions do exist. In the poetic books, for example, an overflow line was sometimes added as the 49th line. This seems to have been done in order to prevent a short overflow line as the first line of a new page (seven times) or of a new column (once).28 Almost all examples are found in scribe D’s section of Psalms: only one example occurs in the text copied by scribe A. In the prose books unclear pricking or ruling could be the reason for a deviating number of lines. Thus in scribe A’s section, the central folding sheet of quire 39 (OT 39–40) by scribe A has 47 lines per column (but see below, page 42, for the second folio of this same quire), quire 57 (starting OT 120) by scribe B contains six deviating columns (46, 47, or 49 lines) spread over four folios, and in quire 88 (starting NT 118) by scribe A, only the first two folios and the last one are regular; all five of the other folios have one or more columns with 47 lines. The one complete column on folio 85.6 (NT 99), at the end of the Pauline corpus, contains also 47 lines and may have to do with the upcoming break. A further three instances defy explanation: in scribe B’s section, folio 43.2 (OT 65) has only columns of 47 lines, just as in the first column of folio 45.2 (OT 81) and the fourth column of folio 58.6 (OT 133). The length of a single full prose line is roughly 13 letters, and that of a full poetic line is about 27 letters. If a line turns out to be shorter than usual, the empty space can be filled up by adding the wedge shaped filling mark ( : ). The optical effect of this is that it gives a more even alignment of the right margin of a column. Its use is concentrated particularly in the work of scribe D and occurs mostly in the prose sections. The use of red ink All the superscriptions to the individual psalms, occurrences of the words diayalma and allhlouia, and numeration of the individual psalms are written in a reddish ink described by Milne and Skeat as vermilion.29 Outside Psalms, the red ink is used also for the rubrics of Song of Songs and in the Eusebian apparatus in the Gospels. A few times it is also used to add some flourishing touches to the coronis.30 Milne and Skeat Examples folio 60.1 (OT 142), col. 2; folio 60.4 (OT 145), col. 2; folio 60.6 (OT 147), col. 2 and col. 4; folio 60.8 (OT 149), col. 4; folio 62.2 (OT 159), col. 4 (for folio 62.2 [OT 159], col. 3 see below page 38); folio 69.7 (OT 220), col. 1 and col. 2. 29 Milne-Skeat, 34. 30 Because of the fading of the original ink and the subsequent retracing, Vaticanus is difficult to describe. Some of the section numbering reveals a reddish 28

38

Non-Textual Characteristics

“suppose that the writer of the text inserted the titles—and with them the numeration, diayalmata and paragraphi—as he went along; otherwise he could scarcely have calculated so exactly the space required, sometimes four, five, or even six lines in length.”31 They also claim that the red ink in scribe A’s section has been retraced, probably by scribe D. The first of these statements is probably incorrect. The argument that scribes would not have been able to calculate the required space is dubious; we will see below that space was calculated over much greater distances than just a couple of lines. Moreover, positive evidence exists indicating that the scribes simply left the space open: a few times a mistake was made in the process. The first such mistake is found on the first page of Psalms: scribe D adds diayalma in red ink to a line on which a word is already written in normal ink.32 In three other places on the same page diayalma receives its own line and this instance forms the only exception in the whole book of Psalms. The same scribe may have forgotten to skip a line for diayalma at the very beginning of folio 62.2 (OT 159), column 4 and added the word later as the 49th line to the previous column. The two most informative irregularities are found in the closing stages of the book and are both by scribe A. Psalm 148 remains without number and without superscription, yet the line for the superscription is left blank.33 On the same page, the transition between Ps 150 and 151 has a remarkable layout. The lines written in red look like this:





allhlouia outosoyalmosidiografoseisda+d kaiexwqentouariqmou oteemonoma chsen twgoliad

allhlouia on line 2 of the column occupies a single line; it is the closing line of Ps 150. The scribe then continues on line 3 with a line of 28 letters, but decreases the number of letters for lines 4–7. This conscious ink under the retracing, but in Psalms I could not find anything comparable. Vermilion is used extensively in Alexandrinus and probably also in the palimpsest Ephraemi Syri Rescriptus, where the absence of the first lines of each Bible book is probably due to the ink being of a different type, which was easily washed away. 31 Milne-Skeat, 35. 32 Folio 59.1 (OT 134), line 2.21, after the word autou. 33 Folio 63.8 (OT 173), col. 1.

Scribes

39

reduction of the number of letters per line is apparently intended to fill up all lines up to line 7. The two explanations are that either the scribe forgot to leave a line blank between the last line of Ps 150 and the superscription to Ps 151 (line 3), or that he had made a miscalculation in how many lines were needed for the red ink section. In either case, it seems that the text in normal ink had been written before the scribe filled in the blanks with red ink. All these examples taken together indicate that instead of the red ink lines being written as the scribe went along, they were added afterwards. Whether this happened at the end of a daily session, when a quire was finished or when the book was completed, is impossible to tell.

SCRIBES Overview of the work of the scribes The identification of the earliest corrections with the scribal hands which are responsible for the main text was one of the most important results of Milne and Skeat’s research. The A and B corrections of text written by scribe A are attributed to scribes A and D, the A corrections in text by scribe D are his own, and likewise the A corrections in scribe B’s text are made by that scribe.34 Thus, scribe A and D have some doings together, but scribe B’s work does not contain a textual contribution by any of the other scribes. The impression might arise that scribe B worked completely unconnected with the other two scribes but this is only partly true. Not only did he add the supplementary apparatus such as titles and page headings to his own sections, but occasionally he contributed such supplementary material to sections written by scribe A and scribe D as well. The following table gives the order of the contributions to the main text by each scribe and the quire numbers for their sections. Within the prophetic section written by scribe B, seven quires are lost. It seems reasonable, though, to suppose that they were written by the same scribe. However, with regard to the first part of the codex very little can be said. The fragments from Genesis are scribe D, quire 12 from Numbers is scribe A, but this leaves large parts of text unallocated. The Contributions of Each Scribe35 Milne-Skeat, 40–46. Square brackets indicate that the book in question is only partially preserved. A continuous line denotes a lacuna in the manuscript, a broken line means that a 34 35

40

Non-Textual Characteristics Scribe D [Gen]

Scribe A

Scribe B

[Num] [12.1 – 12.7] [1 Chr/ 2 Esd]–Esth 34.8 – 37.3 Tob – Jdt 37.3 – 39.2 1 Macc 39.3 – 41.4 4 Macc 42.1 – 42.4 4 Macc 42.4 – 42.8 Isa – [Lam] 43.1 – 49.8 Joel – Mal 57.1 – 58.6 Pss 59.1 – 62.3 Pss – Job 62.3 – 72.8 Matt – Jude 73.1 – 89.1 Rev 89.1 Rev – Barnabas 89.1 – 91.2 [Hermas] 92.1–[94]

In addition to the contributions to the main text, scribe D is responsible for six leaves in the NT which seem to be cancel leaves, that is, the original sheets written by scribe A were removed from the quire and replaced by sheets rewritten by scribe D. The three replaced sheets are 1) the second sheet of quire 74 (folios 74.2 and 74.7 [NT 10, 15]; Matt 16:9– 18:12 and 24:36–25:21), 2) the fourth, inner sheet of quire 76 (folios 76.4 and 76.5 [NT 28, 29]; Mark 14.54–Luke 1:56), and 3) the third sheet of quire 84 (folios 84.3 and 84.6 [NT 88, 91]; 1 Thess 2:14–5:28 and Heb change of hand occurs at a place where the codex is continuous. The numbers refer to the quire numbering.

41

Scribes

4:16–8:1). For the first of these replacement sheets it has been shown that it was replaced after the addition of the Eusebian apparatus as the numbers are extant on either side of the replaced folios but are absent on the newly inserted leaves;36 the second replacement sheet has the Eusebian apparatus. The third replacement sheet lacks the stichoi number, which is present underneath the subscription of the preceding and following epistles. Having established, then, that multiple scribes were responsible for the production of Sinaiticus, the question that arises is, Why was this done? Part of the answer may be that having multiple scribes working concurrently on the same manuscript means that it can be produced more quickly. But also it may have been done simply to break up the immense task of copying a complete Bible.37 Interaction between scribes A and D38

Change of scribe The end of the historical books section shows a remarkable alternation between scribes A and D, almost as if they were taking turns. Below, we diagram their respective contributions and how these align with the quires; the first row shows the quire numbers, the second shows the scribe, and the third, the book. 35 36 37 38 39 40 41 42 43 Scribe A 2 Esd

Scribe D Esth Tob Jdt

Scribe A 1 Macc

D A 4 Macc

B Isa

The prophetic section by scribe B starts at quire 43 and could have been written before or after the two books of the Maccabees were completed. In either case, scribe B wrote his section independently of the other two scribes. See Scrivener, Plain Introduction, 94 and n. 1. Multiple scribes are also identified in other Greek codices containing both Old and New Testament: two scribes are identified in Vaticanus (Canart, “Notice Paléographique” following Milne-Skeat, 87–89) and more than two in Alexandrinus (Milne-Skeat, 90–93). 38 A slightly expanded version of some elements in this section has appeared in Dirk Jongkind, “One Codex, Three Scribes, and Many Books: Struggles with Space in Codex Sinaiticus,” in New Testament Manuscripts: Their Texts and Their World, ed. Thomas J. Kraus and Tobias Nicklas (Leiden; Boston: Brill, 2006), 121–35. 36 37

42

Non-Textual Characteristics

The transition from Esther to Tobit takes place on the recto of folio 37.3 (OT 22) where scribe A finishes Esther on the second column and scribe D starts Tobit at the third column. Nothing irregular can be seen here, the only notable feature being the change of hand. The next change in scribal hand displays a striking irregularity: Judith is finished by scribe D on the last column of the verso of folio 39.2 (OT 37) after which 1 Maccabees by scribe A starts on the second column of 39.3 (OT 38), leaving the first column of the page blank. There is no other instance in the manuscript where a column is left blank between books unless it serves to facilitate the start of the next book on a new page. The text of 1 Maccabees has a regular layout, but the conclusion of the preceding book shows some interesting irregularities. First, the number of lines per column is reduced; both first columns on the two sides of 39.2 (OT 37) have 46 lines per column, while the remaining five full columns have each 47 lines. These columns ignore the ruling, which is set for 48 lines and is still clearly visible. Therefore, the reduced number of lines cannot be attributed to faint ruling. Secondly, the number of filling signs increases over these last pages. Folio 39.1 recto has 14 diplès, folio 39.1 verso has 29, and folio 39.2 recto has 39 filling signs, plus three places where a double filling sign is used. The final page, folio 39.2 verso, which consists of three and a half columns, has 31 filling signs. Thirdly, the rise in filling signs coincides with a decrease in average line length. The line length drops from 13.6 letters per line—as calculated from the first eight columns of the book—to 12.7 on folio 39.1 verso, and even to as low as 11.7 on the seven complete columns of folio 39.2. Although a drop of one letter per line does not sound impressive, in practice it means that, with the lower average, four to five more lines per column are needed to cover the same text. All these measures together amount to stretching out the text by at least one column. It appears that commencing with the recto of folio 39.2, scribe D made a conscious attempt to stretch out the text of Judith. What pressure was he under to do this? If we take into account the blank column before 1 Maccabees, the most probable explanation is that the first columns of 1 Maccabees had already been written before scribe D had finished Judith. This means that, before scribe A started to write, an estimation was made of how much space was required to contain Judith but that this estimation turned out to be wrong. When scribe D discovered the error, he tried to make amendments for it but succeeded only partially. Milne and Skeat have observed that the superscription to 1 Maccabees is by scribe D, while the text is written by scribe A. They assume that the superscription was added

Scribes

43

later.39 In light of our present reconstruction, it may well be that the title was not added later but earlier than the main text in order to indicate the projected start of the next book. The occurrence of the error makes it possible to reconstruct the procedure followed in this particular transition; one cannot tell if a similar procedure (without an error) was followed at the previous transition from scribe A to scribe D, between Esther and Tobit. The end of the text of 1 Maccabees by scribe A reveals another irregularity in this section of historical books. Four Maccabees, which was started by scribe D, follows 1 Maccabees on a new quire (folio 42.1 [OT 56]). The previous book ends neatly on the last column of quire 41. This is certainly no coincidence; a conscious effort is made to contain 1 Maccabees within quire 41. The length of the quire is adjusted by using only two folded sheets instead of the usual four. The decision to use an adjusted quire size must have been made before the scribe started the second half of the quire, but probably earlier. Again we see that the amount of space needed for a particular book was calculated in advance. A further example that illustrates interaction between scribe A and D is found in quire 42, which contains the text of 4 Maccabees and fits exactly onto a single quire—the text ends on the last column of the last page. The first three and a half folios are written by scribe D, the remaining four and a half folios by scribe A. Again we see evidence that the scribes are fully aware of the amount of space they need for their text: scribe D adds two lines per column from folio 42.3 (OT 58) verso onwards and, in the last two columns of the recto of folio 42.4 (OT 59), he even adds three lines. In order to accommodate these extra lines, the parchment is roughened up below the usual area of writing.40 When scribe A takes over on the verso of folio 42.4 (OT 59), the 50 lines per column are continued till the last column of folio 42.5 (OT 60). The last three folios are written regularly and the book finishes nicely on line 38 of the final column. Without the additional lines on the middle pages of the quire, 4 Maccabees would not have fit into a single quire. The interaction between the scribes can be seen in that scribe D had started an action that would ensure that the book would fit within one quire and that this policy was communicated to and executed further by scribe A. What can be the meaning of the peculiar makeup of both 1 and 4 Maccabees? As 4 Maccabees has to extend up to the prophetic books written by scribe B, it is natural that 4 Maccabees was made to finish within its quire, either because scribe B had started Isaiah already, or else because it 39 40

Milne-Skeat, 33 (table). The separate action of pouncing the extra space is still visible in the original.

44

Non-Textual Characteristics

was felt that such a major break in the nature of the collection of books ought to coincide with a quire break. But why was 1 Maccabees made to end on an irregular quire? One possibility is that while scribe A was writing 1 Maccabees, scribe D had already started with 4 Maccabees, and that, therefore, scribe A had to ensure a smooth fit of 1 Maccabees to the next book. Another possibility is that the historical books were intended to end with 1 Maccabees, and that 4 Maccabees was only added as some sort of afterthought. In that case 1 Maccabees is the original ending of the historical books, at which point a shorter quire would not be strange at all. A third possibility is that the scribes made provision for the inclusion of other books in between 1 and 4 Maccabees. Admittedly, such a hypothesis is highly speculative, but the combination of just these two books of the Maccabees and none of the others is unique in the manuscript tradition. The question why only these two books were included and not the more common 2 Maccabees or 3 Maccabees remains open.41 There are two remaining instances where scribe D and scribe A take over from one another. The first one is found in Psalms. Scribe D wrote the first 26.5 folios, scribe A takes over at Ps 97:3.42 The transition takes place at a page break, just as in 4 Maccabees. The second example is found in the NT part of the manuscript. The recto of folio 89.1 (NT 126) contains the conclusion of Jude by scribe A. On the verso scribe D is responsible for the first 34.5 lines of Revelation. After that, scribe A takes over again and finishes the book.

Correction of each other’s work—cancel leaves As mentioned above, scribe D corrects his own work and that of scribe A, but scribe A limits himself to correcting his own work. We find three places where scribe D replaces whole sheets, which were in all likelihood originally written by scribe A, with sheets of his own, all of them in the NT.43 In the first of these places it is hard to reconstruct what was going on. The two folios of the replaced sheet in Matthew do not deviate in the number of letters per line or in their column length. It may be that the reason for replacement was not a substantial dittography or omission. The first folio of the replaced sheet in Paul contains the ending of 1 Thessalonians, of which the last column occupies 21 lines. If this folio originally contained a large 41 1 and 4 Macc have the subscriptions makkabaikwn a (sic) and makkabaiwn d reflecting awareness of the existence of a series of books of the Maccabees. 42 Ps 97:3, folio 62.3 (OT 160) recto and verso. 43 See above page 40.

Scribes

45

repetition or omission, the effects of smoothing this error out would be absorbed by this partially used column.44 However, it is possible that the problem lay in the second folio of the replaced sheet, which contains text from Hebrews and which shows an average number of letters per column above the normal average, suggesting that an omitted passage had to be reinserted.45 The replacement containing the ending of Mark and the beginning of Luke is extensively discussed in Scribes and Correctors.46 Milne and Skeat demonstrate that this sheet was not rewritten because of an existing problem with the ending of Mark, but because of a problem with the opening of Luke. The so-called longer ending of Mark could never have fit on this sheet. Milne and Skeat suppose that in rewriting the original pages scribe D extends the text of Mark into the tenth column instead of the supposed original nine columns. The crowded first six columns of Luke are the rewriting of the original seven columns minus a substantial leap backwards, which was presumably made in the first version of this sheet. Though Milne and Skeat mention the alternative solution that originally a large passage was omitted, they do not give it much consideration, as they assume that scribe D is stretching out the text of Mark from nine to ten columns. But is this indeed the case? The four columns on the verso of folio 76.4 (NT 28) and the first column of folio 76.5 (NT 29) are indeed stretched out, which is not only confirmed by a superficial visual impression of these two pages but also by a rare instance in which the name “Jesus” is written in full, ihsoun.47 44 Incidentally, a large dittography of ten lines occurred in the last lines of the last column just preceding the replaced folio. 1 Thess 2:13, folio 84.2 (NT 87), line 8.25 edexasqe ... – 34 ... tou qu+ repeated in lines 35–44. This might have been the introduction to more problems on the next folio. See the discussion of this passage in Chapter 4, page 207. 45 The first seven columns vary between 629 and 713 letters. See Milne-Skeat, 11 n. 1. 46 Milne-Skeat, 9–11. The two folios of the replacement sheet are folio 76.4 (NT 28) and 76.5 (NT 29). 47 Folio 76.5 (NT 29), line l.27. The writing of other words in full even though a nomen sacrum is available cannot be used as an argument, as scribe D shows a very restricted use of nomina sacra in his normal text. One could think here of forms such as staurow (folio 76.4 [NT 28], line 5.25 and passim) and israhl (folio 76.4 [NT 28], line 6.37, but compare the same scribe at folio 76.5 [NT 29], line 4.35). Tindall gives as a letter count for the five full columns of Mark on the central opening 598, 566, 605, 598, 560, which are all below the average, put by Tindall

46

Non-Textual Characteristics

This stretching out may well have started on the last line of the recto, where we find a double diplè: ::. However, previous to this point, there is no sign of stretching at all. The fourth column of the recto even contains an extremely high number of letters and none of the other three columns show an effort to reduce the number of letters.48 It is therefore not impossible that, in the original version of the replacement sheet, Mark ended on the second column of folio 76.5 (NT 29), just as it does now.49 In that case, the crowded six columns of the beginning of Luke do not indicate a reduction from seven to six columns, but the reintroduction of a previously omitted section.50 This new solution has the advantage of explaining the normal first four columns of Mark by scribe D. If he was faced with the task of lengthening Mark’s nine columns to ten, these very full first columns of Mark are hard to explain. Only at the end of the fourth column of the recto, does scribe D seem to realise that he needs to stretch the text somehow in order to avoid finishing Mark a column too soon.51

Correction of each other’s work—corrections on the text Even by the admission of Milne and Skeat, it is notoriously hard to identify which corrector is responsible for many of the early scriptorium corrections, and certainty is often impossible to achieve. However, in some cases a positive identification can be made. This is true for the larger additions and for those corrections that display one or more of the between 630 and 640. For the six full columns of Luke on folio 76.5 (NT 29) the figures are 684, 679, 698, 685, 728, and 681. 48 Tindall gives for the four columns on the recto of folio 76.4 (NT 28) the numbers of 630, 651, 628, and 708. The number of 708 letters per column is only surpassed three times in the New Testament, all three in replacement leaves by scribe D: once on folio 76.5 (NT 29) verso. col. 3 (728) and twice in the Paul replacement sheet. C. Tindall, Contributions to the Statistical Study of the Codex Sinaiticus, (Edinburgh: Oliver and Boyd, 1961), 28–29. 49 If the omission caused by homoeoteleuton in Mark 15:47–16:1 originated with scribe D, the text would suddenly have become shorter by 76 letters, amounting to five to six lines of text. 50 This possibility was pointed out to me by Dr Leslie McFall (personal communication). 51 A possible explanation, which was not entertained by Milne and Skeat, is that the exemplar of Sinaiticus looked like Vaticanus: an empty column in between Mark and Luke. The main aim of the rewriting by scribe D could then have been the removal of such an empty column. However, the fact that the stretching of Mark starts late and the columns of Luke are crowded, argue against this scenario.

Scribes

47

characteristic hallmarks of the correctors.52 The smaller group of identifiable corrections can be used to address the question of whether scribes A and D both corrected the whole text, or whether they divided the correction work between themselves. If we formulate the hypothesis that both scribes corrected the whole text, evidence for this would consist of either a correction of one scribe on a correction by the other, or the regular occurrence of corrections by both scribes near to one another, for example in the same column or on the same page. There is a caveat here, and that is that the group of corrections by corrector B that occur mainly in the first part of Matthew should not be considered along these lines. Tischendorf thought that all these corrections came from a hand slightly later than the scriptorium hands, while Milne and Skeat attributed them all to scribe A as part of “an ambitious scheme of correction, annotation, and general improvement of the text.”53 If Milne and Skeat’s explanation is accepted, it carries with it the suggestion that the B corrections are part of a possible third correction phase.54 The first phase is formed by the corrections made by the scribe of the main text whilst copying, and the second phase the subsequent correction by scribe A and scribe D of the text of Matthew. For the sake of the argument, we will leave out the B corrections, which do occur close to a correction by scribe D on the verso of folio 73.2. Corrections by scribe A and by scribe D do occur on the same folio, but examples where both scribes each make corrections on the same page are hard to find.55 An example is folio 77.7 (NT 39), where two corrections

52 Milne-Skeat, 40–46, give an overview of the characteristics and apply these to Lake‘s sample on plate II of the NT facsimile of Sinaiticus. On page 43, in the discussion of the ‘A2 corrections’, one should probably read ‘sixth example’ for ‘fifth example’ as the fifth is clearly scribe D and the sixth belongs to scribe A instead of vice versa. 53 Milne-Skeat, 45. 54 Though I find it hard to disagree with Milne and Skeat’s expert judgement, I am not convinced that the B corrections are by scribe A or D. Milne-Skeat, 37, state that the B corrections “gradually merge into A’s more usual hand” but fail to mention that the kai-ligature of the B corrections, one of the most decisive criteria in establishing the various hands, is different from both scribe A and scribe D. See e.g. folio 73.3 (NT 3), line 2.12, and 4.19 in the outer margin. 55 Of course the corrections made by scribe A on his own text which are made in his capacity as prima manus (corrections in the first “correction phase”) are excluded as well.

48

Non-Textual Characteristics

by scribe A are visible on the recto and one by scribe D on the verso.56 On the recto of the next folio, 77.8 (NT 40), we find a clear example of another correction by scribe D, but on the verso a correction by scribe A.57 Milne and Skeat, in their comments on Tischendorf’s notes, list several examples of corrections by scribe D on corrections made by scribe A. At Titus 1:16 scribe D places the word agaqon above the line and erases the same word supplied by scribe A in the margin.58 Likewise, in Matt 12:33, scribe A writes originally alon and corrects this to agaqon. Scribe D corrects this to kalon.59 In Matt 22:9, the scribe writes udatwn initially and then corrects it to odwn by adding an o and d above the a and t and placing a dot above the first two letters. Scribe D then erased the first four letters and replaced them with the two correct ones, adding also a diplè to fill up empty space at the end of the line.60 In none of these cases is it sure whether scribe A made the initial correction while he was writing the main text or whether he made them as he was going through the text as corrector. The precise role division in correcting the text is difficult to establish. In contrast to the situation in most of the NT, the OT shows extensive sections without any early correction at all. For instance, the few corrections made by the early hands in the book of Psalms can all be attributed to the time when the main text was written, implying that no separate effort was made to correct the text. Division of tasks and space restrictions: Revelation, Barnabas, Hermas As we have seen above in the transition from Judith (scribe D) to 1 Maccabees (scribe A), the codex was not written exactly in the order in which it was finally put together. Also, the sections by scribe B are selfcontained with regard to both quire makeup and additional apparatus, albeit this scribe was involved in adding the apparatus to sections written by We know that the correction between columns 2 and 3 is by scribe A because of the kai-ligature, and the correction below column 4 by scribe A because of the arrows. The arrows are also decisive on the verso to identify scribe D. 57 Folio 77.8 (NT 40). That scribe D made the correction between column 3 and 4 on the recto, is almost certain because of the characteristic x. The correction on the verso in the lower margin is by scribe A. For the latter see Milne-Skeat, figure 13. 58 Titus 1:16, folio 85.5 (NT 98), line 4.4. 59 Matt 12:33, folio 73.7 (NT 7), line 4.14. 60 Matt 22:9, folio 74.5 (NT 13), line 4.18. 56

Scribes

49

scribe A. With this in mind, it is worthwhile having a closer look at the situation surrounding the conclusion of Revelation and the following book, the letter of Barnabas. The last two quires of the NT that scribe A wrote are irregular in both the number of sheets and the order in which the sheets are folded together. Quire 90 (NT 134–139) consists of only three folded sheets, quire 91 (NT 140–141) of a single sheet. Additionally, quire 90 is irregular in the way the sheets are placed on top of one another: the rule that is followed elsewhere in the manuscript of facing flesh with flesh and hair with hair side is not followed in the first and last opening of the quire: folio 90.1 verso and 90.6 recto show the hair side, but folio 90.2 recto and 90.5 verso show the flesh side. Also, in Scribes and Correctors it is claimed that in this quire each sheet is ruled individually, which is not according to the customary practice. The transition from Revelation to Barnabas takes place on the recto of the second folio; Barnabas finishes on the last column of the last page of quire 90. Quire 90 Barnabas Revelation f-h . f-h . h-f | f-h . h-f . h-f

Quire 91

f-h | h-f

How can the aberrant quire makeup be explained? If we assume that quire 90 started off as a regular quire, one of the sheets must have been removed, and the obvious candidate would be the second sheet of the original quire, which would explain the current violation of the flesh-hair, hair-flesh order. The individual ruling of each sheet cannot be explained by supposing an originally regular quire, but it may be that the data in Scribes and Correctors are slightly erroneous here. The ruling of a regular quire would imply that the first, outer sheet, and the inner sheet are ruled individually, and that the second and third sheet are ruled on the two openings that show the flesh side of these sheets (namely on folio 2 verso–3 recto and folio 5 verso–6 recto). In the case of quire 90 this would mean that folio 1 plus 6, and 3 plus 4 are ruled as a pair, while folio 2 and 5 have a ruling that would have been completed by folios of the missing sheet. It appeared on

50

Non-Textual Characteristics

personal inspection that the ruling on this quire is very hard to distinguish. However, the pricking in the lower margin is, with some effort, still visible. The first three folios have the pricking at exactly the same position and likewise this is the case with the second three folios.61 This means that this quire was pricked as a single quire and this leaves the possibility open that it was also ruled as a single quire.62 Quire 91 is a single sheet and contains the ending of Barnabas, leaving the last column of the sheet open. The next quire is regular and has the text of Hermas by scribe B. Remarkably, quires 90 and 91 taken together would form a normal quire, begging the question of what might have gone wrong in the original quire 90 that needed mending by the present situation. The text of quire 90 as it now stands does not show any gaps; the “missing” sheet cannot have been taken out after it was completed. But if we suppose that it was taken out after the second folio was nearly completed, it would have contained the last column of Revelation and the first columns of Barnabas. As we have seen above, the breaks between books or book sections were often indicated by starting on a new side of the folio. If that was what scribe A did originally, leaving three columns blank on the hypothetical second folio, he would have run into problems with the ending of Barnabas as this book would need one or two columns of a new quire. If we follow this line of thought further, scribe A must have realised this before he started to pen the third folio, but after he had written a substantial part on the verso of the second folio (so that washing out was no longer a viable option). However, scribe B had started, or even already finished, the next book and it was scribe A’s task to come up with a decent fit. Instead of allowing for a quire containing only a few columns of text, he took the second sheet out, on which he had made the false start for Barnabas, and rewrote the ending of Revelation and the beginning of Barnabas on the original third sheet—now the second—without leaving three columns blank just after Revelation. This explains the distorted fleshhair order, as well as the shorter quire length of quire 90. It was only afterwards that he realised that instead of simply taking the sheet out, he

The pricking in the lower margin is not executed horizontally but is somewhat sloping down. On folio 90.6 verso one should not confuse the real pricking with the marks left by the imprint of the pricking of the facing page. 62 Milne-Skeat, 75, 78, do not claim that each sheet was pricked individually, only that each sheet was ruled individually. Possibly the new conservation assessment of Sinaiticus carried out in the British Library will give the final answer on this issue. 61

Scribes

51

should have replaced it and therefore added the sheet he needed as an extra quire. The above reconstruction is of course purely hypothetical and alternative explanations are possible, such as that scribe A was using up some loose sheets at the end of his task and compiled a deviating quire at the end on the wrong estimation that six folios would be enough to contain Barnabas. Whatever the true reason behind these deviating quires, it is apparent that in the final result, the emphasis of the scribe is not on making a statement about the different (or equal) canonical statuses of Revelation and Barnabas, but on the problem of how to fit the texts into the available space. Supplementary apparatus: superscriptions and running titles The table in Chapter 1, page 14 gives the assessment by Milne and Skeat of the contributions of each scribe to the additional apparatus. Below we will list the patterns and exceptions with regard to the running titles (tituli). Unless otherwise stated, the titles appear on the recto side of a folio. Scribe A provided the titles esdras b on folios 35.1 (OT 4) and 35.3 (OT 6), but scribe B provided the next four titles on 35.5 (OT 8), 35.7 (OT 10), 36.1 (OT 12), and 36.3 (OT 14). The two titles of Esther, like the main text, are again by scribe A (36.7 [OT 18], 38.1 [OT 20]), but for the next book, Tobit, it is scribe B who provides the first two running titles for the text by scribe D (37.5 [OT 24], 37.7 [OT 26]), who himself does the last one (38.1 [OT 28]) and the ones for his next book, Judith (38.3, 38.4, 38.5, 38.7, 39.1, 39.2 [OT 30, 31, 32, 34, 36, 37]).63 After this, 1 Maccabees, by scribe A, remains without tituli; it is scribe D who provides the superscription. Four Maccabees has only one running title above scribe D’s section (42.2 [OT 57]), written—just as the superscription was—by scribe D himself, but we find none in scribe A’s section of the same book. In the prophetic books, scribe B supplies both the running titles and the superscriptions to his own text. Over Isaiah and Jeremiah he places the running titles above the first, third, fifth, and seventh folios of a quire, which is only the recto sides that are also flesh sides of the parchment. The On folio 38.6 (OT 33) a titulus may be present but then only very vague, folio 38.8 (OT 35) is the folio recovered from a bookbinding. The state of this fragment is such that it is not possible to determine the presence or absence of a titulus. Beneshevich does not mention a titulus in his edition of this fragment. V.N. Beneshevich, Catalogus Codicum Manuscriptorum Graecorum qui in Monasterio Sanctae Catherinae in Monte Sina Asservantur (St Petersburg: 1911 [reprint Hildesheim: Olms, 1965]), 639–42. 63

52

Non-Textual Characteristics

Dodekapropheton is without running titles, the exception being the vague and mistaken aggeos above the first opening of Zechariah (58.1 [OT 128]). The sections of Psalms by scribe D and scribe A are both without running titles and superscriptions. In Proverbs, which is from the hand of scribe A, who also added the superscription—though later, according to Milne and Skeat—scribe B adds the first titulus (64.3 [OT 176]), scribe A himself adds the next two (64.5 [OT 178], 64.7 [OT 180]), after which scribe B again provides the running titles on the next folios (65.1, 65.3, 65.5, 65.7 [OT 182, 184, 186, 188]) and also for Ecclesiastes (66.1 [OT 190], 66.3 [OT 192]), for which scribe A had provided the superscription. Scribe A writes both the superscription and the one titulus for Song of Songs (66.7 [OT 196]). Scribe B provides the running titles for Wisdom, Sirach, and Job in the same regular fashion as those for Isaiah and Jeremiah: on flesh sides that are also recto sides. The main text and the superscriptions are all by scribe A, though again the superscriptions for Wisdom and Sirach are added later. In the NT the situation is not less confusing. Scribe A adds the titles himself above the text of the four Gospels but introduces a novelty. Instead of writing the titulus solely on the recto of a folio, it is split up over the two pages that form the opening of a book so that on left side kata is written, and on the right side the name of the Gospel. In Matthew the first full opening has kata maqqaion on both pages (73.1 verso and 73.2 recto [NT 1–2]) but from then on the split title is repeated on each opening, both flesh and hair, except on 73.7 verso–73.8 recto [NT 7–8] and 74.7 verso–74.8 recto [NT 15–16], where it is missing both times, and on 73.8 verso–74.1 recto, where page 73.8 verso [NT 8] has kata, and the facing page kata maqqaion. Tituli by scribe A are present on the first replacement sheet by scribe D (74.2 [NT 10]) and on the recto of the second (74.7 [NT 15]), but not on the opening showing the verso of this folio. In Mark there are no running titles on the first opening, which does already contain the proper superscription on the left page where the book starts. The remainder of the book has running titles only on those openings that show the flesh side, the last being on the replacement folio at the transition from Mark to Luke. On the second folio of the replacement sheet, we find the running title markon by scribe A and the superscription kata loukan by scribe D. The pattern of placing the split titles on flesh openings is continued without exception throughout Luke and most of John. Only at the last opening of John, where the book

Scribes

53

ends on the right hand column of a hair side opening, do we find an “extra” running title across the two pages. The superscription and running titles for Romans are supplied by scribe D on each recto page; the main text comes from the hand of scribe A. The title pros rwmaious is not split up. Scribe D also adds the titles for the first part of 1 Corinthians, but on 82.1 (NT 70) and 82.2 (NT 71), it is scribe A himself who adds the titulus pros korinqious instead of scribe D’s pros korinqious a, as before and after. Scribe A also adds the superscription (but later, according to Milne and Skeat). Scribe D continues with the running titles till 83.6 (NT 83), the end of Ephesians. The next folio does not have a titulus because of the presence of a proper superscription, but 83.8, 84.1, and 84.2 (NT 85–87) have tituli by scribe A (Philippians and Colossians), who also provides the superscriptions to these letters. No running title appears on the only opportunity for one in 1 Thessalonians (84.3 [NT 88]), which is the first folio of the replacement sheet in quire 84. Hebrews has a running title on each right hand page by scribe A but, again, not on the second folio of the replacement sheet (84.6 [NT 91]). Folio 85.2 (NT 95) opens with 1 Timothy and has the superscription pros timoqeos a, as well as a running title without the numeral, both by scribe A. The next folio, 85.3 (NT 96), has a titulus by scribe B, while 85 (NT 97.4) does not show a titulus at all. The next two folios do not have one either, but in these two cases one would not expect one. On the first page of Acts the superscription is positioned in the place of the running title and is attributed to scribe B, who added also the tituli on the right hand page of each flesh opening (except the last incomplete page 88.1). James lacks both a superscription and running title. Milne and Skeat have noticed the peculiar order in which the superscriptions to the numbered epistles arose (except 1 and 2 Corinthians). Initially, the titles to Thessalonians, Timothy, Peter and John consisted only of a numeral, and only later were the words of the full title added. Likewise, scribe A simply wrote a as titulus over 1 Peter and then emended it later. Scribe B mistakenly added the running title petrou b over 1 John (88.7 [NT 124]). Scribe A added the full title epistolh petrou a and epistolh iwannou a to the first letter. For the second and third letter only the name of the author was used, petrou b or iwannou b, iwannou g. Jude has a running title by scribe B (89.1 [NT 126]), the superscription is by scribe A. In Revelation the title is by scribe D, who wrote most of the first column. All the pages on the flesh openings have the split title apokaluyeis iwannou by scribe B, although folio

54

Non-Textual Characteristics

89.8 verso (NT 133) misses apokaluyeis, and so the facing page shows just iwannou. Scribe B also provided the tituli for Barnabas, while the superscription came from scribe A. The two quires on which Barnabas was written are irregular in their size and in the order of the flesh and hair sides, yet we only find tituli on the recto of the flesh sides with the exception of the last opening where we find a titulus on the recto of a hair side, so that both folios of quire 91 have a titulus. The following observations can be made on the basis of this survey: First, the superscription to a book is most often written by the scribe who wrote the main text that followed the superscription. Clear examples of a close relation between main text and superscription are Revelation, where the superscription and only the first 36 lines are by scribe D, and 4 Maccabees, where the superscription and the first three and a half folios of the book are by scribe D. That a superscription was often added later is shown by the fact that in seven cases another scribe provided a superscription for a book written by scribe A.64 No one added a superscription to Psalms or James. Milne and Skeat distinguish between those superscriptions written by the scribe of the text at the same time as the main text and those added later; the present author, however, did not find any data that would support such a fine distinction. Secondly, the addition of the tituli was often done in very small installments. The frequency with which the hands change is remarkable: scribe B added the last four tituli to 2 Esdras over a stretch of 8 folios, next scribe A added the two tituli to Esther, then scribe B added the first two to Tobit, after which scribe D added the last one to that book. One Maccabees did not receive any and scribe D only placed one titulus over the section of 4 Maccabees that he himself wrote. In Proverbs the first titulus is by scribe B, the next two by scribe A, after which, scribe B continues again. In Song of Songs, scribe A again adds a single running title, while the remainder of the poetic section is completed by scribe B. In the NT, scribe A adds two running titles in 1 Corinthians amid a long stretch of titles by scribe D. Thirdly, the running titles were added later than the main text, yet not always at the same stage of subsequent editing of the codex. The three replacement sheets in the NT do not provide a consistent picture. The replacement sheet in Matthew, which lacks the Eusebian apparatus, and in Mark/Luke, which contains the Eusebian apparatus, both have running titles by scribe A, who also provided the other running titles in these books. 64 Scribe D wrote the superscriptions for 1 Macc (though on page 43 we raised the possibility that this was done in order to indicate the starting point for scribe A), Mark, Romans, 2 Corinthians, Galatians, Ephesians; scribe B for Acts.

Conclusions

55

In this case the titles were apparently added after the replacement. The absence of kata on the verso of the second folio of the replacement sheet of Matthew was not caused by the folio being replaced, but because scribe A missed this opening of two pages altogether. In the replacement sheet in Paul we do not find the running titles, although based on the folios before and after one would have expected a titulus on both folios of this sheet. This means that here the tituli had been added before the replacement. Folio 82.4 recto (NT 73) shows that the running title was added after the substantial correction of an omission in 1 Cor 13 by scribe D on the text of scribe A; the titulus has been forced to the left as its usual place was already partly taken by the correction. Fourthly, scribe B is the most consistent scribe with regard to adding running titles only on flesh sides, with just a couple of exceptions. Scribe A follows the flesh-side rule in most of the Gospels, Matthew being the exception with a running title on each opening, as happens also in the Pauline letters. Scribe D adds tituli on each recto page in Paul; in 4 Maccabees he chose a hair side for the only running title in that book. Fifthly, scribe A is not consistent in his handling of numbers in superscriptions and tituli. In the two tituli he provides for 1 Corinthians, the simple pros korinqious is written, in contrast with the tituli written by scribe D before and after. The superscriptions to the other numbered epistles initially consisted simply of the numeral, which was later lengthened to contain the full title of the letter. The same holds true for the titulus for 1 Peter, while the two tituli for 1 Timothy, one by scribe A and one by scribe B, contain only the addressee: pros timoqeon. Though the use of just the numeral may represent a first phase in adding tituli, after which an editorial decision was made to change this to the full title, it may also be that the numerals were added concurrently with the main text. It is difficult to imagine that the scribe intended the simple numeral to be the definitive version of the superscription and the running title. It seems more likely that it functioned as an aid to facilitate the editing process.

CONCLUSIONS The scribes In this chapter we have built upon the work of Milne and Skeat in Scribes and Correctors with regard to the number of hands and their respective contributions in the manuscript. Reflecting upon some of the new observations made in this chapter and the already available evidence, the following remarks can be made.

56

Non-Textual Characteristics

First, scribe B’s sections stand out from the work of the other scribes not only in that he copies and corrects the text, and adds the additional apparatus himself, but also in that his sections stand apart with regard to the physical makeup of the codex. Both beginnings of his two sections coincide with a new quire. Before these new quires there are signs of squeezing the text in the work of the scribe who writes the preceding text. Secondly, the other two scribes, scribes A and D, work more closely together, as is seen especially in the books from 2 Esdras to 4 Maccabees, where they frequently alternate. Instead of writing long, uninterrupted sections, these two scribes write only one or two books at a time, or, in the case of 4 Maccabees, only half a book, before the other scribe takes over. Here both scribes work together in a concerted effort to keep the book within the limits of the quire. The transition of Judith to 1 Maccabees shows that the scribes did not wait for one another, but that before Judith was finished, the next scribe commenced 1 Maccabees. In Psalms the sections of these scribes are considerably larger; the section of Psalms by scribe A forms, together with the remaining poetic books and most of the NT, an uninterrupted stretch of a little more than 25 quires. Thirdly, scribe B does not limit himself strictly to his own sections; he contributes the occasional superscription or running title to text written by the other two scribes. Fourthly, in the NT part of the manuscript, we see all three scribes at work. Almost all the main text is written by scribe A; he also adds part of the additional apparatus. Scribe D writes a small part of text of Revelation, rewrites the three replacement sheets, adds superscriptions and tituli, and makes a considerable amount of corrections on the text. Scribe B adds a considerable number of the tituli. When we consider the contributions and consistency of the work of each scribe individually, scribe B is the most consistent. He adds running titles to flesh pages only and all the quires he uses are of regular length—the last quire of the Dodekapropheton was reduced to six folios by cutting two leaves out, not by using only three sheets. He does, however, make two mistakes in the running titles of Zechariah and 1 John by placing the name of the previous book above the next book, something neither of the other scribes ever does. He corrects the text he has written himself and adds the supplementary apparatus. This last task he also executes at times for text written by the other scribes, though he never corrects their text. Scribe A writes the majority of the surviving text of the codex and does a bit of everything else. He adds some running titles, part of the other supplementary apparatus, and makes a considerable number of corrections

Conclusions

57

to his own text. He does not produce many folios that would not contain a contribution by one or more of the other scribes, either in the form of tituli or corrections. Scribe D contributes the least text of the three scribes, at least as far as the codex has been preserved. He is active in supplying tituli and corrections to scribe A’s work and is responsible for the three cancel leaves in the NT. Three times he starts off a book which he then leaves to scribe A to finish. The production of the manuscript It is clear that no single, fixed procedure was followed in the production of Sinaiticus. The way in which the writing of the main text was divided up between the three scribes seems to betray a number of ad hoc decisions and attempts to cover up previous mistakes. Also their responsibilities seem to vary from place to place. The order of the various stages is also not the same at each place. Consider the order as can be seen in the three replacement sheets in the NT:

58

Non-Textual Characteristics

Order of Contributions on the Replacement Sheets 72.2–74.7 76.4–76.5 84.3–84.6 (Matt) (Mark / Luke) (1 Thess / Heb) Main text - scribe A Main text - scribe A Main text - scribe A Eusebian apparatus scribe D Running titles scribe A Stichoi numbering underneath subscription - scribe A Correction Correction Correction (replacement of sheet) (replacement of sheet) - (replacement of sheet) - scribe D scribe D - scribe D Eusebian apparatus scribe D Running titles Running titles scribe A scribe A The order of the various elements can easily be established by comparing what is present and not present on the replacement sheet in relation to what precedes and follows. The first and second replacement sheets differ in the presence of the Eusebian apparatus on the second. The third replacement sheet lacks the tituli, which were added on the other replacement sheets as the final stage. Additionally, two other relevant examples should be taken into consideration. The first is the correction on 73.6 recto (NT 6), which shows the Eusebian numbers added to a correction placed below the column, indicating that the Eusebian apparatus was added after the correction. The second is folio 82.4 recto (NT 73), where scribe D has supplied a long omission by scribe A above the column. Later, when scribe D added the running title, he placed it further to the left than usual. The irregularities in the insertion of the words written in red ink helped us to establish that these lines were initially left open and only filled in afterwards. The fact that sometimes a particular scribe is responsible for only two consecutive running titles invites speculation on the possibility that this is an indication of the amount of text that was produced in a single day. Could it be that at the end of a day another scribe went over the folios and added

Conclusions

59

the tituli? Though this cannot be proven, we can safely conclude that the editing sometimes took place in a very fragmentary manner, which inevitably would result in a great number of inconsistencies. Concluding observations The first impression one gets when looking at Sinaiticus is one of regularity. The four columns per page in the prose sections, the regular column length, and the controlled, uniform bookhand used throughout the manuscript give the codex a very united look. If creating such an impression of unity was the aim of the scribes, they certainly succeeded. However, under the surface a lot is going on: Three scribes with varying tasks and responsibilities are at times seen struggling to contain books within a quire. They divide their tasks in ways that do not always work out very well and evidence an absence of fixed procedures for correcting and editing. This sense of inconsistency is reinforced by the different levels of attention given to various parts of the manuscript. The NT is well looked after with, at times, frequent corrections and even replacement sheets, but parts of the OT remain virtually uncorrected except for those corrections made by the scribe while writing. Some of the irregularities found in Sinaiticus would remain part and parcel of manuscript production for ages to come. Until the late Middle Ages scribes often squeezed text in to remain within the boundaries of a quire, and tasks were divided among more than one, and sometimes even many scribes. Yet, the way in which Sinaiticus was produced betrays a low level of standardisation in regard to the production and composition of this large codex. In contrast, the bookhand of all three scribes is well formed, regular, and fairly standardised among the three.

3 NOMINA SACRA, LIGATURES, ITACISMS, TEXT-DIVISIONS

INTRODUCTION In the previous chapter we have established that the nature of the interrelatedness of the three scribes is complex and not always fixed. Given this situation, we will now look into specific areas of scribal behaviour in order to reflect on the individual freedom of the scribes. Is it possible to be specific about scribal behaviour at the individual level and to what extent can we expect a scribe to be consistent? The scribes of Sinaiticus had a variety of scribal tools at their disposal to shape the actual form of the text without changing the text that they transmitted. On a textual level it makes no difference whether the scribe uses a ligature for kai or whether he wrote this word out in full. The situation becomes slightly more complicated with the nomina sacra. Did it make any difference to the scribe whether he wrote ihsous or is+ ? And what about the nomina sacra for common words such as pathr, anqrwpos or kurios? Did the use of a nomen sacrum amount to an interpretation of the text? When we move on to consider the insertion of paragraph breaks, the issue of interpretation of the text becomes even more apparent. The insertion of a paragraph break does not change the actual text that is transcribed—the same letters are copied. However, by inserting a paragraph, the scribe could potentially introduce his understanding of certain passages. The question is, of course, whether any of such concerns affected the scribes of Sinaiticus and, if so, whether they affected the individual scribes in the same way. This chapter contains a series of studies on different aspects of the written text, without entering into the discussion of any actual textual variants. Most of the phenomena we will discuss here are part of ancient (Christian) scribal tradition, such as the use of nomina sacra, the kai-ligature, and the inclusion of the Eusebian apparatus. Others, such as itacisms and 61

62

Nomina Sacra, Ligatures, Itacisms, Text-divisions

scribe-determined paragraph divisions, may relate more to the language ability and comprehension of the individual scribe. We will attempt to describe these phenomena at the level of the individual scribe and finally reflect upon the collective results of the individual studies.

NOMINA SACRA The history of research into the Christian phenomenon of the nomina sacra is well known and need not be repeated here in all of its details.1 Much effort has been put into finding the origin and earliest development of the system of contractions for the 15 sacred names. Traube, and most other scholars after him, have assumed a core group of four nomina sacra to which other nouns were added later. These four core members are qeos, kurios, cristos, and ihsous.2 Hurtado describes how opinions have been divided on which of these four nouns was the starting point of the whole system.3 Attention has also been given to a description of the development of the system through the centuries and its application in specific manuscripts.4 It is this latter line of research to which this section seeks to make a contribution. It has been shown by all the major studies on the later development of the system that individual manuscripts are often not consistent in their 1 The most recent overview with extensive bibliography can be found in L.W. Hurtado, “The Origin of the Nomina Sacra: A Proposal”, Journal of Biblical Literature 117 (1998): 655–73. 2 See L. Traube, Nomina Sacra: Versuch einer Geschichte der christlichen Kurzung (München: C.H. Beck’sche Verlag., 1907), 37–38; G. Rudberg, Neutestamentlicher Text und Nomina Sacra (Uppsala, Leipzig: A.-B. Akademiska Bokhandeln, Harrassowitz, 1915), 58–59 (Rudberg expresses uncertainty whether or not pathr and pneuma should also be included in the core group); A.H.R.E. Paap, Nomina Sacra in the Greek Papyri of the First Five Centuries AD (1959), 121; Hurtado, “Origin”, 655–57. 3 Hurtado, 664–71. 4 See Traube‘s listing of the evidence of the major manuscripts, which is often incomplete. Paap presents a chronological overview of the usage of nomina sacra in 421 biblical and non-biblical papyri, split into sacral and profane use. Though the presentation of the evidence follows the format used by Traube, Paap improves on Traube by giving the actual numbers for the plene and contracted forms of each word. The difference in sacral and profane use is subdivided into four different categories in the work of O’Callaghan, Nomina sacra in papyris Graecis saeculi III neotestamentariis (Rome: Biblical Institute Press, 1970); idem, “‘Nominum Sacrorum’ elenchus in Graecis Novi Testamenti papyris a saeculo IV usque ad VIII”, Studia Papyrologica 10 (1971): 99–122.

Nomina Sacra

63

application of nomina sacra. This led the pioneer of the study of nomina sacra, Ludwig Traube, to conjecture that Sinaiticus must have been written by at least six different hands. Traube came to this conclusion starting from the four hands identified by Tischendorf and dividing scribe A and Scribe B into two hands each. The first hand labelled scribe A is responsible for the OT, the second hand for the NT. Scribe B’s work was divided into a hand for Isaiah and a hand for the remainder. The tacit assumption of Traube’s approach is that each of these scribes is internally consistent in his application of the nomina sacra to the text. Though Traube admits the possibility of different exemplars, he seems to assume that a scribe will show internal consistency.5 In his study of Codex Bezae, David Parker found that the nomina sacra had been subjected in the ancestry of Bezae “to revisions of varying degrees of thoroughness and according to various sets of rules.”6 In the final layer, that of the actual scribe of Bezae, Parker notes that the scribe is unaffected by contemporary Greek practice in having only a limited number of contracted nouns in the text. However, the scribe does show awareness of contemporary Latin practice. Thus, Parker explains the current situation in Bezae as the combined result of the practice of the scribe and the differences in the exemplars of the various books,7 though he adds that “a study which seeks too much information from the nomina sacra is of questionable value.” The studies of Traube and Parker present the main questions to which we will try to find an answer. What can we learn about the individual scribe, and do the nomina sacra in Sinaiticus provide us with a possible window on the ancestry of Sinaiticus? First of all we will seek to describe the situation as found in Sinaiticus, both in terms of the forms that were used and the frequency of their use. We will put particular emphasis on the distinctive traits of each scribe and on the level of internal consistency shown by each of them. At the same time, this will also result in an evaluation of Traube’s attempt to identify scribes by the use of nomina sacra. For this section a transcript was used comprising 1 Chronicles, Judith, 4 Maccabees, XII ‘Es gehen unsere Unterscheidungen, denen die Faksimile und nicht die Originale zugrunde gelegt werden mussten, vielleicht zu sehr gerade von dem Unterschied im Gebrauch der Nomina sacra aus, während es an sich möglich ist, dass ein und derselbe Schreiber, von seiner jeweiligen Vorloge abhängig, hierbei ganz verschiedenen Gesetzen folgte’ (Traube, Nomina Sacra, 54). 6 D.C. Parker, Codex Bezae (Cambridge: CUP, 1992), 106. 7 ibid., 106: ‘[the nomina sacra] do betray something of a manuscript’s antecedents and of its scribe’s own habits.’ 5

64

Nomina Sacra, Ligatures, Itacisms, Text-divisions

Prophets (Joel–Haggai), Psalms, Matthew, Mark, Luke, John, Pauline corpus, and Revelation, totalling 175.4 folios. Seven leaves come from scribe B’s work, 36.9 leaves from scribe D, and 131.5 from scribe A. Wider checks have been made on a more general basis. The forms of the nomina sacra The word qeos shows all the normal forms for the singular: q)e, q)s, q)n, q)u and q)w. Besides these five nomina sacra, we find one instance of the genitive plural q&wn (Ps 135:2, folio 63.5 (OT 37), line 4.28): exomologeisqai tw q)w twn q&wn

The same forms are found for kurios (k)e, k)s, k)n, k)u and k&w), including the three occurrences of the genitive plural k&wn.8 Like the genitive plural of qeos, kuriwn is also only employed in the superlative phrase “Lord of lords.” The name “Jesus” needs only three contractions because of its declension ( i)s, i)n, i)u ). Additionally, two stray forms are found as well. The nomen sacrum i&iu in Rev 1:9 looks like a scribal slip but does not seem to have been corrected in the scriptorium; the form i&hu is found in Rev 22:20.9 In the text covered by our sample we have not found any other forms of the nomena sacra of ihsous.10 The fourth word in the core group of nomina sacra is cristos. All the regular two-letter contractions of the singular are found (c)e, c)s, c)n, c)u, c)w), as well as a single instance of the three letter form of the genitive (c&ru) in Rom 7:4.11 The nomina sacra for pneuma and its compounds make for an extensive list. For the singular noun the following forms are found, p&na, p&ns, and p&ni. For this noun, however, the full range of its plural forms 8 Ps 135:3, folio 63.5 (OT 37), line 4.30 exomologeisqai tw k)w twn k&wn. Rev 17:14, folio 89.7 (NT 132), line 6.18 kai to arnion nikhsi autous oti k)s k&wn estin. Rev 19:16, folio 89.8 (NT 133), line 6.16 basileus basilewn kai k)s k&wn. 9 Rev 22:20, folio 90.2 (NT 135), line 1.10. 10 In 1 Cor 1:1 (folio 81.8 [NT 68], line 7.1) the prima manus - according to Tischendorf - corrected an alpha into an iota in order to form i)u. This does not mean that a)u has an independent existence as a nomen sacrum. 11 Rom 7:4, folio 81.3 (NT 64), line 5.30.

Nomina Sacra

65

is present as well, pn+ata, pn+atwn, pn+asin and pn+asi. The adjective pneumatikos can also be contracted, pn+ikos, pn+ikon, pn+ikh, pn+ikhs, pn+ikoi, pn+ikois, pn+ikwn, pn+ikais, pn+aka, as can the adverb pn+ikws. In all these cases, the crossbar that indicates the nomen sacrum is placed only over the first three letters of the contraction; the suffix is not contracted. The contraction proper is always kept together on a single line, but the suffix can be broken off at the end of a line and written on the next such as in the following example, where pn_ika is written over a line break (1 Cor 14:1, folio 82.4 (NT 73), line 4.24–25): zhloute de ta pn_i ka mallon de i

uios has contracted forms only in the singular, u)e, u)s, u)n, and u)u. In our sample we have not found contracted forms of the dative uiw, which would differ only minimally from the nomen sacrum u)w. There is possibly an exception to this last observation in wsanna tw uu+w da+d.12 The word u&u)w has a crossbar and so must have been intended as a nomen sacrum. This unusual form may be explained as a genitive u)u corrected to a dative by simply adding the correct letter after the second upsilon. In normal cases this letter would be expuncted,13 but in this particular instance, the crossbar might have intervened. The noun anqrwpos is in possession of a full set of contractions, an+e, an+o+s, an+o+n, an+o+u, an+w, an+o+i, an+o+us, an+wn + , an+oi+ s. Tischendorf prints an+o once,14 but as this happens at the end of a line, one should probably read an+o( with the final supralinear stroke indicating a nu subsumed into the crossbar of the nomen sacrum. A later corrector changed the reading into ano+_n and also touched up the crossbar. Another irregularity is an+n+w where, probably, the scribe swapped the last two letters and intended an+wn + .15 In Hermas the 16 vocative an+pe + is found. Both the singular and plural of ouranos can be presented by a contraction, ou+n+os, ou+n+on, ou+n+ou, ou+n+w, ou+n+oi, ou+n+ous, ou+nw + n, ou+no + is. Even though the contraction can reach up to six Matt 21:9, folio 74.4 (NT 12), line 5.42. This method of correcting is regularly used by scribe A. 14 Mark 7:20, folio 75.6 (NT 22), line 7.17. 15 Matt 23:13, folio 74.6 (NT 14), line 1.20. 16 Folio 92.6 (NT 147), line 1.26. 12 13

66

Nomina Sacra, Ligatures, Itacisms, Text-divisions

letters, it is always kept on a single line and is never broken off, as is also true for anqrwpos. With contractions of ouranos, one finds occasionally two crossbars over a single contraction (e.g., Job 35:5, folio 72.5 (OT 242), line 4.11): anableyon eis ton ou_no_n kai ide

The three indeclinable Jewish names dauid, ierousalhm17 and israhl are also treated as nomina sacra. The first of these has only one form, da+d,18 the other two have more forms, il_m, ilh+m, ihl+m, ieh+lm, iel+hm, ih_m for Jerusalem, and ih_l and is_l for Israel. In Heb 11:22 the one-off form ish+l occurs;19 in Joel we find also ish+m for ierousalhm.20 A superficial skimming of Isaiah and Jeremiah yielded in addition the form iel+m.21 All the forms with ie could be seen as showing the first two syllables of ierousalhm (the i and e each form a syllable) but are more likely explained as itacistic variations of the corresponding iforms. The noun mhthr can be contracted in each of the four main cases, mh+r, mr_a, mr_s and mr_i, the same is true for pathr, ph_r, pr_a, pr_s, pr_i.22 Also the genitive plural prw+n is found. In Isa 3:6 scribe B has pro+s for the genitive pateros;23 in Jer 3:18, he uses pra+s for the accusative plural. A little later, in 3:25, we find the rather curious pat+er (with crossbar) for the nominative plural pateres.24 Here the nomen sacrum is not formed by contraction but rather by suspension. Another form for the nominative plural is per+s.25 The word swthr has contractions in the singular, sw+r, st_s and st_i, and should not be confused with another contraction for 17 Interestingly, in the manuscript tradition only the transcribed form ierousalhm could be contracted while the Hellenised form taV iJerosovluma is almost always written plene. 18 In the subscription to Psalms we find the unique form da+a. 19 Heb 11:22, folio 84.8 (NT 93), line 6.45. 20 Joel 4:6, folio 57.2 (OT 121), line 3.3. 21 Isa 28:14, folio 44.4 (OT 75), line 4.45. 22 Isa 63:16, folio 46.1 (OT 88), line 7.29 has ph+r with the first two letters combined; see also below Combining letters, page 84. 23 Isa 3:6, folio 43.2 (OT 65), line 1.24. 24 Jer 3:18, folio 46.4 (OT 91), line 8.25 and Jer 3:25, folio 46.5 (OT 92), line 1.33. 25 Isa 64:11, folio 46.2 (OT 89), line 1.42.

67

Nomina Sacra

staurow which occurs in Sinaiticus, str+qh + . We find another contraction for staurow (-es‚ -wqh), in which the letters t and r are combined in a single sign called a staurogram.26 To my knowledge the noun stauros is never contractend. Incidentally, the staurogram as an independent symbol occurs also in the coronis of Jeremiah and in the coronis and subscription of Isaiah. In the book of Revelation, the expression alfa kai w occurs three times. Twice the omega is treated as a nomen sacrum by adding a crossbar over the letter.27 Though w+ in itself is no contraction, it is interesting to see that, to the scribe, the title could be regarded as a true nomen sacrum. Along similar lines, the name qeg+re + i in the fourth vision of Hermas can be explained.28 Relative frequency of the contracted versus the uncontracted forms Not all words that can be written as a nomen sacrum are always written as such, and it is at this point that there is room for further analysis. The following table gives a summary for the analysed folios: Noun qeos kurios ihsous cristos pneuma pneumatikos uios29 anqrwpos ouranos

Number of times as NS 1653 1623 799 468 326 24 157 179 82

Number of times plene 27 29 5 2 4 180 435 245

NS % 98 98 99 99.6 99 100 47 29 25

26 Mark 15:15, folio 76.4 (NT 28), line 4.23; John 19:20 in a correction by an original scribe at folio 80.5 (NT 60), col 1 (with staurogram); Rev 11:8, folio 89.5 (NT 130), line 3.2-3 (with staurogram, broken by line division). For more on the early Christian use of the staurogram see L.H. Hurtado, “The Earliest Evidence of an Emerging Christian Material and Visual Culture”, in Text and Artifact in the Religions of Mediterranean Antiquity, ed. S.G. Wilson and M. Desjardins (Canadian Corporations for Studies in Religion, 2000), 271–88. 27 Rev 1:8, folio 89.1 (NT 126), line 6.8; Rev 21:6, folio 90.1 (NT 134), line 2.40; Rev 22:13, folio 90.1 (NT 134), line 7.48. 28 Folio 92.6 (NT 147), line 2.8. 29 Only the singular is considered as no plural nomina sacra of uios occur.

68

Nomina Sacra, Ligatures, Itacisms, Text-divisions

Noun dauid ierousalhm israhl mhthr pathr swthr staurow

Number of times as NS 179 51 177 33 109 11 2

Number of times plene 80 39 40 82 276 19 4130

NS % 69 57 82 33 29 37 5

From the outset it is clear that a group of five words (six if pneumatikos is seen as a word in its own right) is almost exclusively written as a nomen sacrum, qeos, kurios, ihsous, cristos, and pneuma. Though we saw above that scholars have assumed an original core group of four nouns, in Sinaiticus the core group consists of five words, including those four and pneuma. For this group, the most pertinent question is whether an explanation exists for the exceptions—those instances where the nomina sacra are not used. A second group is formed by the words uios, anqrwpos, ouranos, mhthr, pathr, and swthr. Though ouranos will refer only to the region above the earth (whether or not as the abode of God), the other words of this group can refer to either a very mundane person or to someone connected with the sacred realm. The percentage of words written as nomen sacrum differs from word to word but lies within the range of 25–50%. Here the question is to what extent individual scribal judgement influences the choice of a nomen sacrum. Thirdly, the group of the three Jewish names dauid, israhl, and ierousalhm form a natural unit. The overall preference is to write these names in their contracted form, the percentages being 69%, 57%, and 82% respectively. This means that a substantial number of times the word was still written plene. It is unlikely that a difference in sacral and profane meaning plays a role, which leaves scribal habits as the only explanation left. This, however, begs the question of whether the scribe reproduced the tradition as he found it in his exemplar or whether he imposed his own preference on the final form of the written text.

This number is not based on counting in the manuscript but on a concordance search of the NA27 text. 30

Nomina Sacra

69

Lastly, the verb staurow is contracted only twice in the main text of our sample and once in a correction. The only noteworthy feature is the absence of the nomen sacrum. As regards qeos, almost all occurrences of the plene form are in the plural, which are, as a rule, never contracted (the exception being the single instance of qw+n noted above, page 64), apparently because in the plural the referent cannot be the one true God. Three times the singular noun remains uncontracted. The first occurs in the final stage of Judith, which we saw in the previous chapter was deliberately stretched out by scribe D;31 the second has no explanation but occurs also in Judith;32 and finally, in Ps 41:5 the scribe had to use an extra overflow line for just two words: tou qeou.33 It may well be that in this case the nomen sacrum was written out in full to ensure that the line would contain a more substantial amount of text. The 29 uncontracted instances of kurios contain only six plurals, which are, like the plural of qeos, never contracted except three times in k)s kw+n. Eight uncontracted forms of the singular occur in Matt 25:18– 26, the Parable of the Talents, all referring to the master. The uncontracted form occurs nowhere else in the Gospels. In Judith, the uncontracted form occurs a further 14 times (13 times in a profane sense, once in a sacral meaning. This last instance occurs in a stretched out section),34 over against 42 contracted forms (all referring to God). In 4 Maccabees we find one uncontracted form and one contracted form (neither of them referring to God). With one exception, all the singular plene forms of kurios do not refer directly to a member of the Godhead. As will be shown below, the reverse is certainly not true, a nomen sacrum is no guarantee that a sacral referent is in view. The plene form of ihsous occurs three times in Haggai, referring to the high priest Joshua, and once in Col 4:11, referring to Jesus Justus. The only time the uncontracted name is used while referring to Jesus of Nazareth is in Mark 16:6, and this is in a section that is likely to have been stretched out.35 Of the four uncontracted forms of pneuma, one occurs in the stretched section of Judith.36 Two of them occur on folio 60.5 (OT 146). As Jdt 16:18, folio 39.2 (OT 37), line 6.33. See above, page 42. Jdt 6:8, folio 38.3 (OT 30), line 3.27. 33 Ps 41:5, folio 60.3 (OT 144), line 4.16. 34 Jdt 16:12, folio 39.2 (OT 37), line 5.27. 35 Chapter 2, page 45. 36 Jdt 16:14, folio 39.2 (OT 37), line 5.39. 31 32

70

Nomina Sacra, Ligatures, Itacisms, Text-divisions

with the plene form of qeos only two folios earlier, both these plene forms occur with only one other word in an overflow line. This reinforces our suggestion that the plene form is used in such instances for visual effect. The last plene form occurs on the replacement folio in 1 Thessalonians.37 As the instance at the end of Judith is the fourth occurrence of an uncontracted form of a word that is normally written as a nomen sacrum in a stretched section, it looks as if considerations of space do play a minor role in the decision of whether or not to use a nomen sacrum. The distribution of the nomen sacrum for uios differs markedly from book to book. In 1 Chronicles only one contraction is found over against 41 uncontracted forms,38 and a similar pattern is seen in Psalms. In the Pauline corpus, the ratio shifts more towards the contracted forms but is still predominantly in favour of the uncontracted forms (13 nomina sacra, 27 uncontracted forms). In the Gospels the contractions are preferred: uios

Matt Mark Luke John

Number of times as NS 43 26 38 37

Number of times plene 30 3 26 14

NS % 59 90 59 73

The problem is that, unlike the uncontracted forms of the first group, there does not seem to be a relation between referent and the use of a nomen sacrum. Of the 73 uncontracted forms in the Gospels, 51 refer directly to Jesus; all the others have a non-sacral referent. The distribution for anqrwpos shows a different pattern: anqrwpos

1 Chr Jdt 4 Macc Pss XII Prophets Matt Mark Luke John

Number of times Number of as NS times plene 1 1 8 3 4 9 95 6 7 89 21 33 16 8 90 2 58

NS % 0 11 43 9 46 81 67 8 3

37 Ps 76:4, folio 60.5 (OT 146), line 1.38; Ps 77:8, folio 60.5 (OT 146), line 3.17; 1 Thess 4:8, folio 84.3 (NT 88), line 4.16. 38 1 Chr 19:1, folio 35.4 (OT 7), line 5.16.

71

Nomina Sacra anqrwpos

Pauline corpus Rom 1 Cor 2 Cor Gal Eph Phil Col 1 Thess 2 Thess Heb 1 Tim 2 Tim Titus Rev

Number of times as NS 10 4 2 1 1 1 1

18

Number of times plene 128 24 29 7 13 8 3 7 5 2 10 10 5 5 7

NS % 7

72

The general pattern seems to be that anqrwpos is written plene, with the occasional contraction thrown in. However, five clear exceptions occur: 4 Maccabees and the Dodekapropheton, though both with a limited number of instances, and Matthew, Mark and Revelation. Interestingly, the pattern for anqrwpos in the Gospels does not bear any relation to the pattern for uios. The phrase uios tou anqrwpou warrants special treatment because of its consistent reference to Jesus and the various possibilities of contracting parts of the phrase. The following statistics take, because of the limited contribution of scribe D to the Gospels, only the work by scribe A into account: Scribe A Matt Mark Luke John Total

uios tou anqrwpou 1 5 2 8

u)s tou an+o+u 14 7 1 2 24

u)s tou anqrwpou 2 7 17 7 33

uios tou an+o+u 1 1 2

18 14 24 11 67

The emphasis in Matthew is on the double contracted form, while in Luke and John the form with only uios contracted dominates. Mark holds a middle position. In each Gospel uios is contracted rather than

72

Nomina Sacra, Ligatures, Itacisms, Text-divisions

uncontracted. If we compare the number of contractions within the phrase with the number of contractions outside, it appears that in all four Gospels the percentage of contracted forms of uios in this phrase is higher than outside the phrase: uios as NS (scribe A) Part of the phrase uios tou anqrwpou Elsewhere

Matt

Mark

Luke

John

89%

100%

86%

82%

62%

83%

53%

69%

A similar table can be created for anqrwpos: anqrwpos as NS (scribe A) Part of the phrase uios tou anqrwpou Elsewhere

Matt

Mark

Luke

John

83%

50%

8%

18%

90%

79%

7%

0%

For anqrwpos, no substantial difference exists between the proportion of the nomina sacra used in and outside of the phrase in Matthew and Luke. In John, the only two occurrences of a contracted anqrwpos in the Gospel both occur in uios tou anqrwpou.39 In Mark, a difference exists; here anqrwpos is less often contracted in the phrase than outside the phrase. In summary, it seems that the phrase uios tou anqrwpou has some influence on the use of nomina sacra, though the influence is not present in each Gospel, nor is the influence similar in each Gospel. The phrase contains proportionally more contracted forms of uios in Matthew and Luke than in other contexts and relatively fewer contracted forms of anqrwpos in Mark, but in John it contains all of the contracted forms of anqrwpos. For ouranos the numbers fluctuate also: ouranos

1 Chr Jdt 4 Macc Pss XII Prophets

Number of times as NS

Number of times plene 2 6 4 79 9

NS % 0 0 0 0 0

John 3:14, folio 79.2 (NT 49), line 3.39 and John 13:31, folio 80.2 (NT 57), line 2.12. 39

73

Nomina Sacra ouranos

Matt Mark Luke John Pauline Corpus Rev

Number of times as NS 36 4 1 1 40

Number of times plene 41 11 34 18 21 10

NS % 47 26 3 5 0 80

These numbers cannot be used to say anything about the text outside our sample. For example, many instances of a contracted ouranos and anqrwpos occur in Job, so that it is certainly not the case that some sort of OT/NT divide exists. The following figures for pathr and mhthr confirm the pattern that was seen with regard to anqrwpos and ouranos in Matthew/Mark and Luke/John; the latter two Gospels have considerably fewer nomina sacra than the first group.

1 Chr Jdt 4 Macc Pss Matt Mark Luke John Pauline corpus Rev

pathr NS plene 8 2 7 4 7 1 21 58 5 11 6 2 43 5 121 16 56

NS % 0 22 36 5 92 65 4 4 22

4

67

2

mhthr NS

2 2 16 11 2

plene

NS %

1 30 10 7 4 14 9 6

0 6 17 70 73 13 0 0

1

0

The data for dauid show that only in Psalms is there a substantial deviation from the rule to write dauid as a nomen sacrum. ierousalhm is written plene mostly in Judith and Luke; the other books prefer the contracted form.

1 Chr Jdt 4 Macc Pss

dauid NS 105

plene

NS % 100

2 28

1 75

67 27

ierousalhm NS plene 8 4 11

NS % 100 26

14

82

3

74

Nomina Sacra, Ligatures, Itacisms, Text-divisions dauid NS

XII Prophets Matt Mark Luke John Pauline corpus Rev

16 7 12 2 3

ierousalhm NS plene 9 3

NS % 75

94 100 92 100 100

2

100

100

2

plene

NS %

1 1

3

5

21

19

7

1

88 100

Differences between the scribes

The data In what follows, we will attempt to establish individual patterns for the use of nomina sacra by each scribe. Thus, we will not concentrate on the (legitimate) question, “To what extent do the scribes copy the nomina sacra from their exemplar?” but on how each individual scribe applied the nomina sacra to their text. The scribal use of nomina sacra was already prevalent, and the scribes knew about the various possibilities within this system of contractions and adapted them to a certain degree towards their own preferences. All the way through, the possibility that the actual exemplar did exercise influence upon the final result must be left open. Let us turn first to the two books written by two different scribes, 4 Maccabees and Psalms. 4 Maccabbees Scribe D Scribe A NS plene NS plene qeos kurios pneuma anqrwpos ouranos uios dauid israhl pathr mhthr

6 1 1

1

2 1 1 1 3 2

33 1 3 2

1 1 8 2

2 3 2

4 28

Though 4 Maccabees is not a very large work, a few differences between the two scribes emerge. Scribe D does not use the contraction of

75

Nomina Sacra

kurios in the phrase oti kurios estin twn paqwn o logismos.40 In a comparable phrase, scribe A uses the nomen sacrum: monos gar sofos kai andreios estin o twn paqwn k)s. Obviously, in both cases the referent is non-sacral. The name dauid is written both ways by scribe D, but only in its contracted form by scribe A (note however that the numbers are extremely small). The treatment of pathr and mhthr is notably different. While scribe D writes these consistently plene, scribe A tends to mix these forms, even though none of the pathr references is to the heavenly Father. The noun anqrwpos is treated similarly by both scribes, using both the contracted and plene form. The data from 4 Maccabees can be compared with the other book both scribes worked on, the book of Psalms: Psalms qeos kurios pneuma cristos ouranos anqrwpos uios dauid ierousalhm israhl mhthr pathr swthr

Scribe D NS 380 477 13 9 6 2 1 16 1 1 1

plene 11 2 45 67 13 74 3 16 7 14 8

Scribe A NS 60 332 16 2 3 1 26 13 29 1

plene 1 1

34 28 2 1

3 7

Given the nature of the text, it is not surprising to find no plene forms of kurios, the plural in Ps 122:2 being the exception. Twice in scribe D’s section, however, we do find pneuma uncontracted in overflow lines, and once qeos,41 while we find no exception to the contracted forms in scribe A’s section. The situation for anqrwpos in Psalms is not dissimilar from that in 4 Maccabees. The one marked difference between the scribes in Psalms is found in their treatment of the three Jewish names. Scribe D strongly prefers the 4 Macc 2:7, folio 42.1 (OT 56), line 7.7. Ps 76:4, folio 60.5 (OT 146), line 1.38; Ps 77:8, folio 60.5 (OT 146), line 3.17; Ps 41:5, folio 60.3 (OT 144), line 4.16. 40 41

76

Nomina Sacra, Ligatures, Itacisms, Text-divisions

plene form for dauid (always spelled as daueid) and regarding ierousalhm and israhl the ratio is roughly 1:1. Scribe A, on the other hand, shows a remarkable consistency, preferring the nomina sacra with only one exception.42 When scribe D uses a contraction for israhl he prefers is_l (16 times) and uses il_m for the ierousalhm contraction.43 For israhl, scribe A alternates between the contractions ih_l (6 times plus once in a correction) and is_l (23 times), and consistently uses ilh+m for ierousalhm. Unlike 4 Maccabees, no difference exists between the two scribes with regard to pathr and mhthr. Comparing the work of Scribes A and D in the NT is somewhat problematic in that the NT sections by scribe D were written as replacement leaves, and that notions such as stretching or condensing the text have influenced scribal practice. The first case is found in Matthew where scribe D wrote the replacement sheet consisting of folio 74.2 and 74.7 (NT 10 and 15): Matt

Scribe A

qeos ihsous cristos kurios anqrwpos ouranos uios pathr mhthr

NS 50 130 14 54 79 36 40 55 16

pl. 1

2 10 30 17 3 7

Scribe D folios 74.2 and 74.7 NS pl. 2 18 3 11 8 10 11 11 1 13 3 2

Scribe D folio 74.2

Scribe D folio 74.7

NS

NS

pl.

9 1

8 6 2 5 1

2 9 1 2

pl.

5 9 8 1

1

First of all the different treatment of kurios should be noted. The two uncontracted forms by scribe A are both plural, the eight uncontracted forms by scribe D are all singular and, as seen above, are all used within the a single section, the Parable of the Talents, where they do not refer to God. Combined with the other uncontracted singular instances of kurios in Judith, it may be said that all the uncontracted singular forms of kurios in the codex are by scribe D and do not—with the exception of a single 42 43

Ps 131:17, folio 63.5 (OT 170), line 2.45. Ps 50:20, folio 60.6 (OT 147), line 2.35.

77

Nomina Sacra

instance in a stretched-out section—refer to the deity. Scribe A, without exception, uses always the contraction, irrespective of the referent. The differences regarding the three Jewish names found in Psalms cannot be confirmed from the Matthew replacement sheet as these do not occur in Scribe D’s section. The difference in treating pathr and mhthr as found in 4 Maccabees but not in Psalms may be present here, with scribe A overwhelmingly preferring the nomen sacrum for pathr. The three uncontracted instances (one of them plural) of scribe A all occur with non-sacral referents, the two plene forms in scribe D’s section are both to God. For anqrwpos, ouranos, and uios a difference between the two scribes is found that was not present in Psalms: scribe A prefers the contracted form for each of these, while scribe D chooses the full form, except for anqrwpos on the first replacement sheet, where scribe D comes closer to scribe A’s normal practice. In the Pauline corpus the figures for scribe D and A are as follows: Paul qeos ihsous cristos kurios pneuma pneumatikos anqrwpos ouranos uios dauid ierousalhm israhl pathr mhthr swthr

Scribe D folios 84.3 and 84.6 NS plene 31 14 8 24 3 1 1 2

6 2 3

1

2

Scribe A NS 628 216 389 434 164 24 15 11 4 16 28 17 8

plene 3 1 2

122 29 24 2 1 54 6 6

There is hardly any difference discernible between the work of the two scribes, the only difference might be the treatment of uios, though the numbers are too small to draw any conclusion. Remarkably, we find again the uncontracted pneuma only with scribe D, just as in Psalms.

78

Nomina Sacra, Ligatures, Itacisms, Text-divisions

The last replacement sheet to be discussed is the one containing the ending of Mark and the opening of Luke. Mark qeos ihsous cristos kurios pneuma anqrwpos uios ouranos dauid israhl pathr mhthr staurow

Scribe D folios 76.4 and 76.5 NS plene 5 12 1 4

1

2 1 1 1 1 1

Scribe A NS 44 72 4 16 22 33 24 4 7 2 11 11

plene

14 2 10 1 6 4

1

The only two remarkable features are the occurrence of an uncontracted ihsous in the stretched out closing section of Mark by scribe D and the difference in the treatment of anqrwpos. Luke qeos ihsous cristos kurios pneuma anqrwpos uios ouranos dauid ierousalhm israhl pathr mhthr swthr

Scribe D folio 76.5 NS plene 10 1 13 5 1 5 1 2 2 1 1

1

1 1

Scribe A NS 110 86 13 87 31 7 38 1 11 5 10 1

plene

1 90 21 34 21 42 13 1

Nomina Sacra

79

In Luke, there is possibly a different treatment by the two scribes of anqrwpos, though the sample for scribe D is extremely small, and probably a difference exists in the treatment of uios. No difference exists in treatment of the Jewish names. With pathr, mhthr, and swthr scribe D has relatively more contractions. Here, scribe A uses predominantly the plene forms. If we compare the work of scribe D with his section of Mark on the same replacement sheet, we find more nomina sacra in the condensed section of Luke, and at least one remarkable plene form in Mark. The individual patterns for each scribe There is ample evidence for the conclusion that the individual habits of the scribe have a large influence on the use of the nomina sacra. Already, by setting the data for scribes A and D side by side, some characteristics emerge, with scribe D being more restrictive in the use of nomina sacra and, in some cases, even paying attention to the nature of the referent. A complicating factor is that it is not clear whether both scribes impose, to a certain degree, their preference upon the text and, consequently, act as editors, or whether only one of them does this, while the other transmits the nomina sacra as he finds them in his exemplar. The emphasis on nomina sacra in scribe D’s section of Luke may well be related to the condensed nature of the text. However, the numbers are not high enough to draw firm conclusions.

Scribe A We have seen that scribe A does not differentiate between sacral and nonsacral referents for kurios,, but instead prefers to contract the noun consistently. In the table “Overview of Scribe A” on page 257, the data for scribe A have been collated by book. Besides the “core group” of five words (qeos, ihsous, cristos, kurios, and forms of pneuma), a few other words receive a fairly constant treatment: dauid, israhl, and swthr, though the last one does not occur often. The remaining words in the table receive different treatment in different parts of scribe A’s work. The majority of instances of ouranos in Matthew and the overwhelming majority in Revelation are contracted, everywhere else the plene form is preferred. As a rule, ierousalhm is contracted, except in Luke. mhthr has more contracted than uncontracted forms in Matthew and Mark, but elsewhere the reverse is true. Yet for pathr the contracted form is preferred in 4

80

Nomina Sacra, Ligatures, Itacisms, Text-divisions

Maccabees, Matthew, Mark, and Revelation. Revelation shares with Matthew the feature that for almost all of these words the emphasis is on the contracted form. Mark is very close to these two books, but in this book the use of the nomen sacrum and plene forms tends to be more balanced. If we consider just the four Gospels for the words anqrwpos, ouranos, uios, ierousalhm, pathr, and mhthr, Matthew has the highest percentage of nomina sacra for most of these words, Mark an equal or lesser number, and still lesser figures for Luke and John. It seems as if a sliding scale is visible throughout the four Gospels, with more contractions in Matthew and Mark than later on.

Scribe D As can be seen in the table “Overview of Scribe D” on page 259, this scribe is remarkably consistent in his application of nomina sacra throughout the various books he writes. When we have a look at the three words anqrwpos, ouranos, and uios, it appears that in each book or section written by scribe D, usually the uncontracted form is used, sometimes even exclusively. The one notable exception is anqrwpos in Matthew; here both forms are used about the same number of times (10 contractions and 11 plene forms). The three Jewish names are treated equally throughout scribe D’s work. ierousalhm is contracted in roughly a quarter of all cases, israhl in half of all cases, and dauid only rarely, though the vast majority of instances of dauid occur in just one book, namely Psalms. The words in the next group of words (mhthr, pathr, swthr) occur less often than words from the previous two groups, which increases the uncertainty of any observed trend. Yet it seems that in the LXX scribe D prefers the plene form above the contractions, in contrast to the larger number of contractions in his NT sections. In the table presenting the details for scribe D, the numbers for pneuma are also given. The only instances of the plene form of this word occur in sections written by scribe D. The same holds true for kurios, of which we have seen that all but one of the 23 plene singular forms have a secular referent.44 With this in mind, it is worthwhile to also consider the fragments of Genesis (OT 1) copied by scribe D. Indeed, here also the scribe distinguishes between Abraham and God as referent of kurios. Abraham is referred to by kurios written plene, God by the nomen sacrum.

44

See above, page 69.

81

Nomina Sacra

However, kurios is the only word for which we find such a consistent distinction. A further question concerns the influence of space considerations. Scribe D had to stretch out the text of Judith, and the last six columns of Mark show signs of stretching, too. The text of Luke that scribe D wrote (the first six columns) had to be squeezed into a limited space. We found that there was a tendency to write nouns in full (even in places where a nomen sacrum could have been used) in stretched sections and that the opposite was true in sections where text needed to be squeezed in.

Scribe B As can be seen in the chart below, scribe B prefers, in our sampled section, the nomen sacrum above the plene form of the word. Interestingly, the double occurrence of ihsous, referring to the high priest Joshua, is treated as a true nomen sacrum. On the other hand, criston, in Hab 3:13, is not contracted; it refers to the people. The only instance of a contracted pathr is the genitive plural prw+n, referring to the fathers of the people.45 For a correct understanding of scribe B’s practice, his wide variety of contractions must be taken into account. Almost all of the unusual forms of ierousalhm are the product of his pen: ish+m, ieh+lm + , iel+hm + , ih_m. As we will see below, this diversity corresponds with his erratic spelling. Scribe B cristos ihsous kurios anqrwpos uios ouranos ierousalhm israhl pathr swthr

45

NS 2 159 6

9 8 1

Joel 1:2, folio 57.1 (OT 120), line 1.31.

plene 1

6 2 6 3

1

82

Nomina Sacra, Ligatures, Itacisms, Text-divisions

Conclusions on the use of nomina sacra in the analysed sample In 4 Maccabees a difference emerged between the usage of the nomen sacrum kurios in the singular by the two scribes who worked on the book. Scribe D did not contract the only instance of kurios in his section, whilst scribe A did. In Matthew, again, scribe D does not contract kurios in the Parable of the Talents on the replacement sheet, while scribe A contracts consistently, irrespective of the sacral status of the referent. As we have seen that in Judith scribe D also distinguished carefully between a sacral referent for kurios and a non-sacral referent, we may take this as a specific feature of scribe D. Also in 4 Maccabees, we noticed a difference in the treatment of pathr and mhthr. Scribe A prefers the contracted form for pathr and every now and then uses the contracted form of mhthr. Scribe D consistently uses the plene form. This difference in 4 Maccabees disappears altogether in Psalms, with both scribes predominantly using the plene forms. In Matthew, scribe A uses the nomen sacrum in the majority of cases. Here, even scribe D has more contractions than plene forms, albeit in a more moderate ratio than scribe A. In Paul there is hardly any noticeable difference, both scribes using two to three times as many plene forms as contractions. Psalms provided us with a marked difference in the treatment of the three Jewish names israhl, ierousalhm, and dauid: scribe D prefers the uncontracted forms, while scribe A prefers the nomen sacrum. Only in Luke can the scribes be compared with one another on this point, scribe D having the nomen sacrum of dauid in one out of two instances and israhl in both occurrences. Scribe A uses contractions in all cases. As scribe D had to squeeze in the text of Luke at this point, it may well be that this had some influence on his practice. In Matthew a stark contrast is seen between both scribes’ treatment of anqrwpos, ouranos, and uios, with scribe A preferring again the contracted forms for each of these. In the case of anqrwpos, he is followed by scribe D to a certain extent but not completely. In Mark, scribe A’s practice is different, but there is still a contrast with scribe D in the treatment of anqrwpos and uios. In Luke, scribe A prefers the contraction for uios, which contrasts with the five uncontracted forms in scribe D’s constrained text. Scribe D does not use a nomen sacrum, withstanding the pressure to save space. It may be of influence that the nomen sacrum of uios does not yield a great space

Nomina Sacra

83

advantage over the plene form. The single instance of a contracted anqrwpos is not enough to make any positive conclusion. In Paul, scribe D uses an uncontracted pneuma, which never occurs in scribe A’s text. In the entire sample that we used for this chapter, all four uncontracted forms of pneuma are limited to the text of scribe D. The system of nomina sacra is not consistently applied in any of the books we have studied in this section. Some of the individual nouns are contracted consistently, but most display a notable variation between contracted and plene forms. Also the system as represented in Sinaiticus is not uniform in all its parts, as demonstrated by the variation between the forms for israhl, ih_l and is_l, and the variety of forms found for ierousalhm. None of the three scribes is internally consistent in the use of contracted words: each of them has words that sometimes are and sometimes are not written as a nomen sacrum. The origin and subsequent use of nomina sacra is connected with reverential notions, but the use of nomina sacra in Sinaiticus is not determined solely by reverence. Often nomina sacra are employed with very mundane referents: pneuma and its derivatives are used both for evil spirits and for the Holy Spirit, and pathr and kurios can also be used in any context. Only in the work of scribe D is there a clear distinction between the reverential and the non-reverential use of kurios. However, this scribe also does not hesitate to write qeos or ihsous plene when it suits his considerations of space. The phrase uios tou anqrwpou is not treated in the same way in the four Gospels. Within all this variety, individual scribal preferences can be detected. This means that to a certain extent the scribes imposed their own preferences upon the presentation of the text. At the same time, we can see a flux within the characteristic patterns of each scribe; a scribe does not always show the same preferences. The unanswered question is whether this fluctuation is simply due to the whim of the scribe or to different practices in their exemplars. In the case of the replacement leaves, the added problem is which, if either, of the two scribes reflects the practice of the exemplar. A satisfactory answer to these questions cannot be found on the basis of the data presented here. If we reassess Traube’s analysis of Sinaiticus, we agree partially with Traube that individual scribes show individual scribal habits in using nomina sacra: the two halves of Psalms and the replacement leaves of the NT do testify to individual patterns. However, in the case of these examples it is impossible to get access to the practice in the exemplar. If we assume that

84

Nomina Sacra, Ligatures, Itacisms, Text-divisions

one of the scribes imposed his own system and the other did not, it still follows that the scribe who imposed his own system was not consistent in doing this throughout all his work. Both scribe A and scribe D show considerable internal fluctuation. This lack of internal consistency makes it impossible to identify different scribes purely on the basis of differences in the patterns of nomina sacra. Milne and Skeat were therefore right to ignore Traube’s proposals and to concentrate instead on the palaeographical argument. However, to a certain degree the individual patterns do lend support to the palaeographical assessment.

LIGATURES The scribal phenomenon of combining adjacent letters into a single grapheme is well known and often discussed in the palaeographical handbooks.46 The ligatures in Sinaiticus can be divided into three groups: 1) combinations of letters that share a vertical stroke; 2) the ligature of mou (‰) and pro(s); and 3) the kai-ligature.47 Instances of these ligatures occur predominantly towards the end of a line, though, with the exception of the second group, not necessarily as the last letter of a line. In corrections these ligatures are more frequently employed.48 Combining letters In order to probe into the scribal habits of the various scribes, we sampled 129 leaves of Sinaiticus for the use of combined letters. In Sinaiticus, the most common combinations are those in which two letters are combined, and especially combinations involving eta (e.g. ©). Far less common are combinations of three letters (e.g. ~), and even more rare are combinations of this type of ligature with the mou-ligature. There is one instance where scribe A, acting as corrector, added a kai-ligature to an See e.g. B.M. Metzger, Manuscripts of the Greek Bible (New York: Oxford University Press, 1981), 30; E.M. Thompson, An Introduction to Greek and Latin Palaeography (Oxford: Clarendon Press, 1912), 57 (though here only the occurrence of this phenomenon in Latin manuscripts is acknowledged). 47 Strictly speaking we could add the staurogram to this list. The occurrences of this ligature in Sinaiticus are listed on page 66 under the nomina sacra for staurow. 48 The data for this section come from the following books: 1 Chr (scribe A), Judith (scribe D), 4 Macc (scribe D and A), XII Prophets: Joel–Haggai (scribe B), Psalms (scribe D and A) and the four gospels (scribe D and A). The pseudofacsimile by Tischendorf is not very useful as it only notes some of the most common combinations, ignoring all the others. 46

85

Ligatures

existing nu, using the last vertical of the nu as the basis for the kappa.49 In John 1:14 an original nh combination was resolved back into two separate letters, apparently by the scribe himself or a scriptorium corrector.50 The alternative form of the letter mu is the so-called Coptic mu (µ).51 The regular mu lends itself to combinations of letters, as the first and last vertical strokes are straight. With the Coptic mu these lines are curved. It is only scribe B who combines this mu with a following eta.52 By far the most common combination is that of n and h, where the last vertical of the nu is combined with the first of eta. Six out of the 126 instances involve a combination over a word break in which the last letter of one word is combined with the first letter of the next word. Other regular combinations occurring more than 10 times are mh and hn (respectively 57 and 47 instances). The frequent occurrence of just these combinations may be due to the nature of the Greek language in which these letters occur relatively often. The hn-combination has an alternative abbreviation at the end of a line in the form of an eta with a crossbar extending to the right (h( ). Prose sample LXX Combination Scribe D (11.3 folios) nh 16 mh 1 mn 3 nhg 2 hn 1 ph hh mn hk 1 hm 1 mh 1 mnh hmn 1

Scribe A (9.4 folios) 17 6

Scribe B (7 folios) 13 7 1

4 1

8 1 1

1

Matt 12:44, folio 73.7 (NT 7), line 6.8. John 1:14, folio 79.1 (NT 48), line 2.12. 51 This type of mu became common in the Coptic alphabet but it does not indicate a link with Egypt. The Coptic mu was used throughout the Greek world (Milne-Skeat, 66). 52 Hab 3:16, folio 57.6 (OT 125), line 1.36. 49 50

86

Nomina Sacra, Ligatures, Itacisms, Text-divisions Prose sample LXX Combination Scribe D (11.3 folios) nhm 1 Total (per folio): 28 (2.5)

Scribe A (9.4 folios)

Scribe B (7 folios)

28 (3.0)

32 (4.6)

The range of combinations of letters differs from scribe to scribe. Scribe D does not have many instances of the mh combination but has the greatest diversity of combinations of three letters. Scribe D and scribe A show a comparable frequency of combined letters; scribe B has the highest frequency. In the Gospels the situation is not much different, but here scribe D’s text is exclusively those of the almost four replacement folios: Gospels Combination nh mh hn ph hm hr hk Total (per folio):

Scribe D (3.9 folios) 6 1 1

1 9 (2.3)

Scribe A (56.5 folios) 53 53 41 26 6 1 1 128 (2.3)

The poetic lines in Psalms demand abbreviation or contraction less often than the prose lines do. The end of a poetic line always forces the start of a new physical line in the manuscript, and when a poetic line runs over into a second or even third line, these overflow lines are often not needed in full. Because there are only two columns of text per page, the actual number of line endings is reduced by half. Given the change in circumstances, one would expect a drastically reduced range and number of ligatures if their main use was that of reducing space. For scribe A this is indeed the case: only five ligatures are found, divided over two letter combinations. Psalms Combination nh n‰

Scribe D (26.5 folios) 19 5

Scribe A (13.9 folios) 2

A possible ligature nh occurs at Matt 1:15, 73.1 (NT 1), 1ine 2.37 but the text is hard to read. 53

87

Ligatures Psalms Combination nhm hn hg hhg hn‰ nm nhn nhk nh‰ mh nhg hk h‰ hhm hng Total (per folio):

Scribe D (26.5 folios) 5 4 4 3 3 3 1 1 1 1 1 1 1 1 1 55 (2.1)

Scribe A (13.9 folios) 3

5 (0.4)

The work of scribe D is surprising in the use of combined letters. When compared with the section of scribe A, scribe D has a higher frequency of ligatures (2.1 per folio) distributed over 17 different combinations of letters. Seven of these combinations include three letters, and four combinations involve the mou-ligature. Two of the combinations do not involve an eta. It seems that, to scribe D, the combined letters came to function primarily as ornaments. mou and pro(s pro s) ligatures The ligatures used for the syllables mou and pros occur as a rule only towards the end of a line. The ligature can stand for a word in itself but equally it is used for a syllable that forms part of a larger word, such as in nomou.54 The mou-ligature can be written using a regular mu (‰) or a mu of the Coptic type (¤). In both cases a short vertical stroke is added to the middle of the mu, representing the upsilon, and a tiny omicron is placed above the letter. erhmous Isa 50:3, folio 45.5 (OT 84), line 2.16; nomou Acts 18:15, folio 87.2 (NT 111), line 5.35; agnismou Acts 21:26, folio 87.4 (NT 113), line 7.10. 54

88

Nomina Sacra, Ligatures, Itacisms, Text-divisions

The pro(s)-ligature is formed by writing the rho through the high horizontal stroke of the pi, while the remaining letters are either suspended or written in full ( pr ). The only instance of the pro(s)-ligature with -os

suspended that I could find was in the running title to 1 Timothy, identified by Milne-Skeat as a superscription by scribe B.55 All the occurrences in the main text of the pro(s)-ligature are also found in the text written by scribe B, but no suspension of the letters -os takes place in any of these instances.56 The mou-ligature is more widespread but still relatively rare: mou-ligature Prose sample LXX Psalms

Gospels

Scribe D 1 (0.1)

Scribe A 4 (0.4)

21 + 10 composite (1.2) -

14 (1.0)

Scribe B 6 (0.9)

6 (0.1)

Scribe B has the highest frequency of the three scribes in the prose sections. Both scribes D and A have a stronger preference for the ligature in the Psalms when compared to their prose sections. The 10 composite forms are those mou-ligatures in combination with the combined letters discussed in the previous paragraph. As with the previous type of ligature, its increased occurrence in Psalms may be explained by aesthetic considerations, shared by both scribes D and A. The kai-ligature kai Milne and Skeat found the form of the kai-ligature one of the clearest signals for each scribal hand.57 This ligature is formed by writing a kappa with a short line attached to the downward arm of the kappa ( Â ), which may be slightly shortened for the occasion. Scribe D has a curved medium length stroke, scribe A a medium, straight stroke at an acute angle, and scribe B has either a very short line, or a line with an angle in it ( k< ). The

Folio 85.3 recto (NT 96). The following list of five examples is probably not complete but is indicative of the low frequency at which this particular ligature occurs. Isa 7:22, folio 43.3 (OT 66), line 8.29 pro- (bata); Isa 26:15, folio 44.2 (OT 73), line 7.8 pros(qhsei); Jer 5:23, folio 46.6 (OT 93), line 2.40 pro- (imon); Jonah 1:14, folio 57.3 (OT 122), line 4.18 pros. 57 Milne-Skeat, 22–23. 55 56

89

Ligatures

difference between the three scribes is fairly consistent and can be seen in their corrections as well. The kai-ligature is, of course, often used as an abbreviation of the copulative particle kai, but can be used for the syllable as such. Examples of this are mainly found in the Pauline letters, mostly in connection with the dikaios word group.58 The ligature is found mostly at the end of a line in prose texts. In poetic texts it is found occasionally a few letters before the end of a line.59 As both examples occur with scribe A and these are the only occurrences of the ligature in scribe A’s part of Psalms, this may indicate that this scribe approaches poetic texts in a different way than he does prose. kai-ligature Prose sample LXX Psalms Gospels

Scribe D 51(4.5) 6 (0.2) 17 (4.4)

Scribe A 39 (4.1) 2 (0.1) 131 (2.3)

Scribe B 41(5.9)

The kai-ligature is rarely used in Psalms, suggesting that it was not regarded by either scribe as a means to decorate the text. The difference in usage between scribes D and A in the Gospels is remarkable and may point to the special character of the replacement leaves, which had to be squeezed into existing material or to a reduced interest by scribe A in the use of these ligatures. Additional counting in Romans and Revelation shows a frequency in Romans that is comparable with that found in the Gospels (1.8 per folio) and a frequency for Revelation comparable to that of the LXX prose section (4.8). Conclusions on the use of ligatures A clear difference between the three scribes shows up in their strategy of combining letters by means of sharing a vertical stroke. In both the Psalms and the LXX prose texts, scribe D uses the widest range of combinations, followed in the prose texts by scribe B, who has the highest frequency. In the Gospels little difference can be found between scribes D and A. The 58 Scribe A: Luke 12:57, folio 77.7 (NT 39), line 7.21; Rom 1:17, folio 81.1 (NT 62), line 2.24; Rom 3:10, folio 81.2 (NT 63), line 1.29; Rom 3:26, folio 81.2 (NT 63), line 3.6 (in kairos); Rom 5:7, folio 81.2 (NT 63), line 7.24; Rom 8:30, folio 81.4 (NT 65), line 3.48; 1 Cor 4:4, folio 81.8 (NT 69), line 8.19; Heb 11:11, folio 84.8 (NT 93), line 5.25 (kairos). Scribe B: folio 92.1 (NT 142), line 8.5. 59 Ps 103:25, folio 62.5 (OT 162), line 1.37; here the kai-ligature is followed on the same line by the word eurucwros. Ps 111:4, folio 62.8 (OT165), line 3.21; in the word dikaios.

90

Nomina Sacra, Ligatures, Itacisms, Text-divisions

function of these combined letters is not primarily to save space. Particularly in Psalms, it seems that scribe D employs this ligature in an ornamental role, combining it regularly with the mou-ligature. In the work of scribe D and scribe A, the mou-ligature occurs more often in poetic text than in prose text. As regards the kai-ligature, it appears that this ligature was not strictly limited to the copula kai but could also be used for the syllable -kai-. Only rarely is the kai-ligature used in a position not at the very end of a line. In contrast to the other two types of ligatures, this ligature is used much less often in the poetic texts. Wherever a comparison was possible, it turned out that scribe B uses any of the three types of ligature more often than scribes D and A.

ORTHOGRAPHIC PATTERNS Introduction In Scribes and Correctors, extensive attention is given to the orthography of the three scribes of Sinaiticus.60 Milne and Skeat describe the main characteristics of each of the three scribes. Scribe D is the most reputable of the three.61 His errors are the substitution of ei for i and i for ei in medial positions, both equally common. Otherwise substitution of i for initial ei is unknown, and final ei is only replaced in the third declension dative singular. Only in iscuei is the final i replaced with ei. A “worse type of phonetic error,” confusing e and ai, is very rare, and if it happens, the substituted form is nearly always a grammatical form of a Greek word. Scribe A is described as markedly inferior to scribe D. Confusion of ei and i is rampant, substitution of i for ei outnumbers the converse by four or five to one, especially in verbal termination. In the third declension dative singular, ei for i is found repeatedly. Confusion of e and ai occurs in all contexts, but in the definite article the correct orthography is most often maintained. Scribe B receives very little praise from Milne and Skeat. A scribe “careless and illiterate” without a “firm visual impression of Greek,” producing forms that are “barbarous and grotesque” with “utter inconsistency.” He produces many forms of third declension accusatives 60 Milne-Skeat, 51–55. See also the catalogue of characteristics in F.H.A. Scrivener, A Full Collation of the Codex Sinaiticus with the Received Text of the New Testament (Cambridge: 1864), liv–lvi. 61 The following description follows Milne-Skeat, 53–55.

91

Orthographic Patterns

with a final -n, phonetic errors defy reckoning, especially confusion of e and ai. Additionally, confusion of liquids and stops occurs frequently, as well as the vulgar omission or insertion of gamma. Omissions of letters and syllables are incredibly common. More curious is the wrong insertion of the consonant in the middle of a word, and also the error of metathesis occurs frequently. In this section we will deal mainly with a select group of orthographic variants: interchanges between ei and i, and between ai and e. We have also noted all the variation between the pairs h and ei, and oi and u. The term most commonly used for this type of spelling variation is “itacism.” Some object to the use of this term as it is not only factually incorrect (not all changes deal with /i/ sounds) but it is also apparently perceived as pejorative.62 Despite its shortcomings, “itacism” is a useful term to distinguish the type of variation dealt with in this section from the more diverse group of orthographic variants discussed in the next chapter. The following data are the result of a collation of 95 folios of Sinaiticus. The portions analysed are the book of Psalms, written by scribes D and A; Joel to Haggai in the Dodekapropheton by scribe B; and the three synoptic Gospels, of which four leaves are by scribe D and the remaining 43 by scribe A. The collation base was the printed edition by Rahlfs for the LXX books and the text of NA27 for the Synoptics. Itacisms per scribe The following table shows the number of itacisms per folio, excluding the itacisms found in proper nouns. The column headings should be read as “i in the collation base substituted in Sinaiticus with ei.” Itacisms per folio XII prophets Scribe B Psalms Scribe D Scribe A Gospels Scribe D Scribe A

Total

i–ei

ei–i

e–ai

ai–e

other

78.2

2.8

33.6

9.2

30.4

2.2

9.2 21.1

3.7 1.8

4.9 13.6

0.2 3.6

0.2 1.4

0.2 0.7

9.4 46

2.8 1.7

4.1 29.5

0 4.3

0.5 10.0

0 0.5

As expected on the basis of the description in Scribes and Correctors, the figures for scribe B are worse than those of any of the others. His See C.C. Caragounis, The Development of Greek and the New Testament (Tübingen: Mohr Siebeck, 2004), 492 and 500–01. 62

92

Nomina Sacra, Ligatures, Itacisms, Text-divisions

confusion of e with ai is even more frequent than that of i and ei. Relatively frequent are the mutual interchange of -h- and -(e)i- (1.4 per folio), where -h- should be often interpreted as having iota subscript,63 followed by the oi–u pair. Combined with his tendency to omit letters, especially intersyllabic g, some words are barely recognisable. So, for example, the form foiin is found in Jonah 1:3, which contains two itacisms (u–oi and ei–i) and misses the g.64 The scribe himself noted and corrected this form. At the other end of the spectrum we find far fewer itacisms in scribe D’s sections in Psalms and the Gospels. Six times the scribe writes the initial syllable of various forms of eidon as idon, which is apparently not counted by Milne and Skeat as an initial ei. The overall pattern of itacisms is roughly the same in both sections. If one allows for the difference in the amount of text between a folio written in the prose format and one written in the poetic format (approximately 5500 against 4000 letters per folio), then the numbers are somewhat higher in the poetic section. The difference may well be due to the fact that in the Gospels, the leaves by scribe D are replacement rather than primary leaves. Three times an (e)i is replaced with a -h-, once an -h- with a -i-, and also we find su instead of soi.65 Scribe A is consistent in having fewer instances of i–ei substitutions than scribe D in both Psalms and the Gospels. However, scribe A has a considerably higher number of ei–i and ai–e substitutions in the Gospels than in Psalms, which begs the question of how consistent this scribe really is. In the “other” group we find substitution of -h- with -(e)i- and vice versa (in total six out of ten), and four times -u- instead of an -oi-. The difference between scribe D and A is well illustrated by the graph “LXX: Itacisms per Folio” on page 260. In Psalms, the transition between the two scribes takes place on folio 62.3 (OT 160), the different constitution of itacistic errors is obvious. The similar graph drawn up for the three synoptic Gospels (Gospels: Itacisms per folio, page 260) shows a surprising result: though the four folios of the replacement sheets by scribe D (74.2; 74.7; 75.4; 75.5) stand out as expected, the first two quires of Matthew’s text show also a significant difference to scribe A’s pattern 63 One may question the validity of including the ei - h/ pair in the section on spelling. Particularly in verbal endings both forms are genuine readings and could be the difference between the subjunctive and the indicative. 64 fugein Jonah 1:3, folio 57.3 (OT 122), line 2.13. 65 This last oi - u change is not necessarily an orthographic variant given the following vocative form kurie; Ps 61:13, folio 60.8 (OT 149), line 4.29.

93

Orthographic Patterns

elsewhere. The number of times ei is replaced with i and ai with e is clearly higher than elsewhere. It is in this stretch of text that scribe A pens nine times the feminine dative plural article as tes and produces on folio 74.6 (NT 14) six times the form ke for the copulative kai.66 On the other hand, the i-ei change occurs significantly less often than elsewhere in the work of scribe A. The deterioration in some departments combined with an improvement in others is hard to explain by variation within the usual pattern of the scribe and leaves us with only two possibilities: a) the orthography of the exemplar has a relatively large influence on the way scribe A copies Matthew, or b) there is a change related to the scribe. Possibility a) would explain the increased number of replacements of ai with e but assumes that scribe A suppressed his normal pattern of writing i for ei and actually produced less of these interchanges than he normally would have. The second possibility, a change related to the scribe, could entail a number of things. One could think of a different way of copying, of reading aloud, or even of a different scribe altogether. This last option is a real possibility and will be examined more closely in the conclusion to this chapter. The different pattern of itacisms in Matthew allows us to draw up a more precise profile of scribe A’s normal spelling pattern as found outside the first two quires of Matthew (73–74). Scribe A Psalms Gospels quires 75–77 Gospels quires 73–74

Total 21.1

i–ei 1.8

ei–i 13.6

e–ai 3.6

ai–e 1.4

other 0.7

34.1

2.1

22.7

4.7

4.2

0.5

70.0

0.6

43.5

3.4

22

0.5

Itacisms in proper nouns The correct spelling of proper nouns contains an element of choice on the part of the editor of the modern text edition. For that reason, this category was singled out in order not to distort the overall pattern. The following table gives the itacisms that occur in proper nouns as compared with the standard text editions, and it will be clear that this yields a different picture.67

I have not found this particular orthographic variant elsewhere in scribe A. I have included the words far(e)isaioi and saddouk(ai/e)oi in this category. 66 67

94

Nomina Sacra, Ligatures, Itacisms, Text-divisions Itacisms in proper nouns XII prophets Scribe B Psalms Scribe D Scribe A Gospels Scribe D Scribe A

i–ei

ei–i

e–ai

ai–e

1.1

0.4

0.3

2.3

1.7 0.1

0 0

0 0

0 0

8.0 1.0

0.3 0.1

0 0.0

0 0.6

The distortion of the pattern without the proper nouns is particularly great for scribe D. In the book of Psalms, this scribe has 35 times daueid instead of dauid, while scribe A normally uses a nomen sacrum. In the Gospels, the high number is mainly due to scribe D’s peilatos (10 times), eleisabet (7 times), and eleias (6 times). Of these three names, only eleisabet is once written without the -ei-;68 the other names are consistently spelled with ei by scribe D. Though scribe D may display a clear preference to write proper nouns with ei rather than i, the same tendency can be seen in the other two scribes as well. Still, a difference between the scribes is clear. It may well be that the modern editors of the Greek biblical text show too strong a preference for the forms with -i- rather than -ei-. In any case, the difference between the practice of the scribes remains and is not affected by modern editorial decisions. Conclusions on itacisms Our short study has broadly confirmed the remarks of Milne and Skeat on orthographic errors. We have sought to quantify the exact differences between the three scribes. The characteristic patterns turned out to be less stable than would have been expected. In Matthew, scribe A shows a remarkable deterioration over the length of two quires. What exactly caused this change remains unclear. Scribes A and B make the error of substituting the longer for the shorter form more often than the reverse for both the i - ei and e - ai pair. Though scribe D also does this, he does so to a lesser degree. His number of changes of ei for i is higher than for both the other scribes. However, overall, he has the least number of examples of orthographic errors in his work. 68

Luke 1:5, folio 76.5 (NT 29), line 3.28

Paragraphing

95

PARAGRAPHING Introduction The study of the devices that are used to delimit textual units has recently received much more attention through the establishment of an ongoing series of publications under the title Delimitation Criticism.69 Though originally conceived as a method to gain access to ancient interpretation of the Hebrew Scriptures, the study of delimitation markers in manuscripts of the NT has also been taken up seriously.70 The most recent description of the way in which paragraphs are used in early Greek biblical manuscripts comes from the hand of Emanuel Tov.71 He demonstrates that the various methods of indicating new sections can all be found in early, pre-Christian biblical manuscripts. In our study we will not reflect upon the implications this has for the development of the paragraphing system and its (now proven to be erroneous) use in the dating of manuscripts.72 The question to which we will seek an answer in this section is whether the marking of the end of one paragraph and the start of the next is something left to the decision of the scribe or whether this is part of a transmitted scribal tradition. Connected with this question is the issue of the meaning the paragraph breaks. Were paragraphs used to indicate the structure of the text? Methods to indicate paragraphs The definition of paragraphs we use in this section is purely a formal one. A paragraph is a section of text marked off by at least two of the following four features: start of a new line, an extruded first letter, a space of at least several letters on a line, or a paragraphus above the first full line of the paragraph. This definition rules out minor breaks in a text indicated simply M.C.A. Korpel and J.M. Oesch, eds. Delimitation Criticism: A New Tool in Biblical Scholarship (Assen: Van Gorcum, 2000) and idem, Studies in Scriptural Unit Division (Assen: Van Gorcum, 2002). 70 E.g. in a paper read by Stanley Porter to the Pericope seminar at the International meeting of the SBL, Groningen 2004. 71 E. Tov, Scribal Practices and Approaches Reflected in the Texts Found in the Judean Desert (Leiden; Boston: Brill, 2004), 303–15. Similar work is done by R.A. Kraft, “The ‘Textual Mechanics’ of Early Jewish LXX/OG Papyri and Fragments”, in The Bible as Book: The Transmission of the Greek Text, ed. S. McKendrick and O.A. O’Sullivan (London: The British Library, 2003), 51–72. 72 The way paragraphs are indicated in Sinaiticus was in the eyes of Sanders an argument for a later dating. See above, Chapter 1, page 20. 69

96

Nomina Sacra, Ligatures, Itacisms, Text-divisions

by a space of one or more letters and breaks merely indicated by interpunction.73 The method most often employed in Sinaiticus is to indicate a new paragraph by starting on a new line that projects slightly into the left margin (ekthesis). This means that often a large part of the previous line remains blank. Manuscripts with broader columns, like Vaticanus and Alexandrinus, prefer to leave only a space of a couple of letters before the next paragraph, after which the first letter of the next line is either projected into the left margin and enlarged (Alexandrinus), or the line is marked with a paragraphus (as in Vaticanus, which occasionally also uses extrusion). In Sinaiticus, when the last line of the previous paragraph has been used up completely, the new paragraph is indicated only by the extruded first line of the new paragraph. In a few cases, Sinaiticus does not leave the remainder of the line blank but continues with the next paragraph after a short space. In Luke we find one example in which we have a space at the paragraph break, and the first new line of the paragraph projecting into the margin.74 In Matthew something similar happens, but instead of the extrusion a paragraphus is used.75 In 4 Maccabees the method of a space on one line and a paragraphus on the next line is used intermittently with the normal usage,76 but the total number of paragraphs is small in this book. In one place in the Minor Prophets it looks as if a new paragraph starts with the nonsense word pria.77 However, this is the result of a rewriting of two erased letters on the previous line and of the first two of this line. In order to facilitate the correction to sapria the scribe projected the first letter of line 45 into the left margin. All the rare cases are exceptions to the overall rule and do not constitute a hierarchy of paragraph divisions. In Vaticanus, a hierarchy of major and minor divisions can be seen in Matthew and to a lesser degree also in Mark and Luke (elsewhere only sporadically). Here, the open paragraph with extrusion and without a paragraphus indicates the major division, the more frequent intralinear open space with paragraphus indicates the minor division. As noted above, the first letter of the paragraph projects into the left margin in most cases, but the length of this projection is by no means fixed. 73 B.M. Metzger (Manuscripts of the Greek Bible, 32) is mistaken when he describes scribe D’s use of the diplè ( : ) as having a role in paragraphing. 74 Luke 19:16, folio 78.3 (NT 43), line 5.20. 75 Matt 1:18, folio 73.1 (NT 1), line 3:19. 76 See e.g. 4 Macc 4:4, folio 42.2 (OT 57), line 4:46. 77 Joel 2:20, folio 57.1 (OT 120), line 7.45.

Paragraphing

97

The most extreme example of the variation within the work of a single scribe is found in the opening folios of the Gospel of Mark. On the first ten pages the letter projects no more than 3 millimetres into the margin, but very often there is hardly any projection at all. In some cases one can only be sure of a new paragraph when the previous line has not been filled out. However, from the third column of the verso of folio 75.7 (NT 23) onwards, the ekthesis becomes grossly exaggerated and amounts to a full letter or more.78 On the verso of folio 75.8 (NT 24) the ekthesis goes back to a more normal length and remains such for the rest of the book. Frequency of new paragraphs The frequency of paragraphs varies from book to book. A graphical overview of the number of paragraphs in a selection of books can be found below on page 262. The highest frequency is found in the Pauline corpus where the folios of Romans contain many paragraphs (folio 81.1–81.7 [NT 62–68]). Folio 81.3 (NT 64), Rom 5:16–8:5, has 62 paragraph starts. For comparison, this section has only 59 verses in the modern versification. At the other end of the spectrum stands folio 88.7 (NT 124), 1 John 2:24–5:5, with only 12 paragraphs. Romans and 1 John differ in style, but style alone is not sufficient to explain the difference. Again, 4 Maccabees is obviously a different genre, and here the paragraphing may have been also influenced by the need to contain this book within a single quire, but still the occurrence of not a single break in the text on folio 43.7 (OT 70) stands out.79 The sample from scribe B (seven folios of the XII Prophets) can hardly be described as representative, given the many book-endings which often leave parts of the last column of a book blank. However, one of the seven folios, 57.3 (OT 122), stands out clearly by its high number of paragraphs. This folio contains the bulk of the book of Jonah, the only narrative book in this prophetic section. The book of Jonah was well known in the patristic period and its familiarity, as well as its different genre, may have contributed to its different treatment. However, it is impossible to tell for this particular book whether the number of paragraphs is due to the exemplar or to the scribe.

78 79

Folio 77.8 (NT 40), line 1.45 has almost a projection of 2 letters. For details on the quire of 4 Maccabees, see Chapter 2, page 43.

98

Nomina Sacra, Ligatures, Itacisms, Text-divisions

Differences within the work of a scribe The number of paragraphs differs widely within the work of a single scribe and even within a book. Part of this may be explained by the nature of the underlying text; some texts lend themselves more easily to multiple breaks than other texts. Assuming for the moment that the scribe is responsible for the insertion of the paragraph breaks, his understanding of the text comes into play as well. It is likely that well known texts received a qualitatively better treatment than obscure or difficult texts. Concentrating on the work of scribe A, it appears that Paul’s letter to the Romans receives the most dense treatment, while Revelation has the least paragraphs per folio in the NT. For example, in the messages to the seven churches only two paragraph breaks are found, the first starting with tw aggelw ths en efesw ekklhsias grayon tade legei ktl. (Rev 2:1), the second paragraph (Rev 2:8) opens with the words tade legei o prwtos kai o escatos in the second of the seven letters, leaving the introduction to this letter inside the previous paragraph.80 The next paragraph starts exactly a full folio later with the agios agios of the four creatures in Rev 4:8.81 We see that here the very obvious signs of delimitation present in the text itself are ignored. Less extreme is the difference between the Pauline corpus and the Catholic Epistles. On the whole, the Catholic Epistles show about half the number of paragraphs per folio found in the Pauline letters. Of the four Gospels, Luke shows the widest distribution: some folios with many paragraphs, others with much less. This is hard to explain on the basis of the nature of the underlying text alone. Four Maccabees is ideal ground to compare scribes D and A, as each wrote about half the text of this book. The following table shows the distribution of the paragraphs over the two scribes: Paragraphs – Paragraphi –

Scribe D (3.5 folios) 12 5

Scribe A (4.5 folios) 6 182

The difference in frequency seems obvious: scribe D has more than double the number of paragraphs compared to scribe A, but neither of them has as high a number of paragraphs per folio when compared with the text of the NT books (with the exception of Revelation). The low number Rev 2:1, folio 89.1 (NT 126), line 8.3; Rev 2:8, folio 89.2 (NT 127), line 1.5. Rev 4:8, folio 89.3 (NT 128), line 1.5. 82 The single paragraphus occurs in the first column written by scribe A. 80 81

Paragraphing

99

of paragraph breaks could be accounted for by the constrictions of space for 4 Maccabees, especially since the reverse phenomenon can be seen in at least one other place. While the first five full folios of Judith all have between four and nine breaks per folio, the last two folios have respectively 17 and 20 paragraphs. We have seen before that this part of Judith was deliberately stretched out. A difference between the two scribes can be seen in the replacement leaves of the NT as well, albeit that scribe D has a lower number in these cases. As for the leaves in Matthew by scribe D, 74.2 and 74.7, the first of the two leaves stands out clearly, the second is closer to the local average. Folio 73.7 73.8 74.1 74.2 74.3 74.4 74.5 74.6 74.7 74.8 75.1 Paragraphs 35 25 34 16 26 33 34 36 28 40 29

The numbers for the second replacement sheet cover the last ten columns of Mark and the first six of Luke and have to be understood in the light of the transition between these books in the middle of folio 76.5 recto (NT 29): Folio 75.7 75.8 76.1 76.2 76.3 76.4 76.5 Paragraphs 39 40 19 20 11 47 5 16 (M) (L)

76.6 76.7 76.8 12 11 11

The folio of Mark written by scribe D clearly has more paragraphs than its immediate context and contains the highest number of all the Gospels. The six columns of Luke on folio 76.5 (NT 29) have a higher frequency than the following leaves but deviate not exceptionally from levels seen elsewhere. These figures confirm the earlier suggestions that folio 76.4 (NT 28) is a folio on which the text is stretched out.83 The last replacement sheet is found as folios 84.3 / 84.6 in the Pauline corpus. Folio84 83.8 84.1 84.2 84.3 84.4 84.5 84.6 84.7 84.8 85.1 85.2 Paragraphs 28 34 35 16 38 37 13 38 35 36 35

In both the replacement leaves we see a reduced number of paragraphs, the two lowest scores in the whole of the Pauline corpus. The first of these two leaves, 84.3 (NT 88), contains the ending of 1 Thessalonians on the last column. The lower number of paragraphs is therefore not necessarily due to space considerations, as extra text would be 83

On page 45 it was argued that this is only true for the second half of this

folio. Folios 83.8; 84.2; 84.3; 84.4 and 85.1 contain the ending of a book and therefore have a column that is not fully filled. 84

100

Nomina Sacra, Ligatures, Itacisms, Text-divisions

absorbed in the final column of the book, but is more likely to be a reflection of either the scribe’s preference or the influence of the exemplar. The quality of paragraph breaks If the paragraph is used to demarcate a smaller or larger portion of text, one would assume that its beginning and end would concur with the natural breaks in the flow of a text. As seen above with the examples from the seven letters of Revelation, paragraph breaks are by no means consistent, and in such instances, paragraphs can hardly be said to be an aid to the reader in seeing the structure of the text. Below we shall discuss examples in which paragraph breaks are inserted at places where they are dubious even when considered on a mere syntactical level.

Scribe B in the prophets The paragraph divisions of the first five chapters of Isaiah were checked in order to find out to what extent these divisions made sense in the immediate syntactical context.85 At least seven of the 51 paragraph breaks are dubious or inexplicable. Isa 1:1, folio 43.1 (OT 64), line 1.7–10 aqam . kai acaz . kai ezekiou . oi ebasileusan ths ioudaias .

The paragraph starts with a relative pronoun following a list of four kings. Lines 9 and 10 contain only a comment on the four kings and hardly merit being cut off from the previous words. A possible explanation for this particular paragraph division is that the relative pronoun oi{ was interpreted as the article oiJ being used as a personal pronoun. 85 The delimitation markers of the first 12 chapters of Isaiah in Sinaiticus were also studied by Wim de Bruin, “Interpreting Delimiters: The Complexity of Text Delimitation in Four Major Septuagint Manuscripts”, in Studies in Scriptural Unit Division (Assen: Van Gorcum, 2002), 66–89. His description of the scribal practice in this part of Sinaiticus is marred by several misunderstandings of what actually is going on in the manuscript. So for example (p. 74) “In this view we have to consider that the text of [Sinaiticus] was written in relatively small columns, which [sic] very limited room for spaces. For this reason a lot of delimitations have even become invisible.” The aim of De Bruin’s paper is to give an overview of the various ways in which units of text were demarcated.

Paragraphing

101

Isa 1:4, folio 43.1 (OT 64), line 1.22–29 ouai eqnos amartwlon . laos plhrhs amartiwn . sperma ponhron uioi anomoi . enkatelipate ton kn+ kai parwrgisate ton agion israhl .

The words sperma ponhron are part of a series of appositions to eqnos amartwlon in line 22 and cannot be seen as the start of a new paragraph. The second paragraph starting at line 28 is the second half of a straightforward parallelism. Isa 1:21, folio 43.1 (OT 64), line 4.5–10 kai alhqias . en h dikaiosunh ekoimhqh en auth . nun de foneute . to argurion umwn adokimon . oi ka-

The contrastive concluding comment nun de foneute has become the start of a new paragraph. Isa 1:26, folio 43.1 (OT 64), line 4.39–43 fanous (v 26) kai episthsw tous kritas sou ws to proteron kai tous sumboulous sou ws to ap archs

Even though the main verb of lines 39–40, episthsw, also governs the second half of the parallelism, which starts on line 42, a paragraph break separating the two has been inserted. Isa 3:3, folio 43.2 (OT 65), line 1.1–8 teron . (v 3) kai penthkontarcon . kai qaumaston sumboulon kai sofon

102

Nomina Sacra, Ligatures, Itacisms, Text-divisions arcitektonan kai suneton akroathn (v 4) kai episthsw neaniskous

The new paragraph starts on line 5 with the last item of a list mentioning things the Lord will remove from Judah. It is almost impossible to construe kai suneton akroathn with what follows. Isa 3:20–21, folio 43.2 (OT 65), line 3.33–39 kai thn sunqesin tou kosmou ths doxhs autwn . kai tous klidwnas kai ta enwtia . kai ta periporfura . kai ta mesopor-

Both paragraphs in these lines occur within the long list of luxury items the Lord will take away. There seems to be no apparent reason for the insertion of paragraph breaks at these particular points. Spot checks further on in the work of scribe B show that the great majority of paragraphs are less awkward—when seen on the level of the sentence—than the examples discussed above, yet it seems that at least at this point in Isaiah, an element of mechanical paragraphing was used, which focused on signal-words such as de and kai.

Scribe A in the Gospels A phenomenon occasionally found with scribe A, is that a paragraph is inserted only on second thought: initially the scribe wrote a line without a paragraph break, then he erased the last letters of this line in order to rewrite them on the next line as the start of a new paragraph. Matt 7:26, folio 73.4 (NT 4), line 4.35–3886 86 The reconstruction of the erasure is on the authority of Tischendorf. Though a close inspection under normal light did not help me, the erased space matches exactly the required space. The notebook by Milne and Skeat (Add. Ms. 78935 A) does not contain a comment on this correction. Another example is found in Luke 9:21, folio 77.4 (NT 36), line 4.25, where after to an erasure is visible. Tischendorf assumed an i, which would be an orthographic variant for the next syllable ei-. No systematic survey of the manuscript has been carried out to collect all the examples.

Paragraphing

103

petran (v 26) kai pas o

This was corrected to: petran {erasure} kai pas o akouwn mou tous logous toutous kai mh poiwn

The length of the erasure in line 36 shows that the correction was carried out before the scribe progressed to writing akouwn. A more confusing instance of the same phenomenon is found in Luke 10. Luke 10:13, folio 77.5 (NT 37), line 3.44–47 ekinh ouai ouai soi corazein ouai soi bhqsaidan oti ei en turw kai

In this example lines 45 and 46 are both extruded into the margin. The first ouai at the end of line 44 has been erased, probably by the original scribe. Instead of having a dittography of ouai, it is much more likely that we have a situation analogous to the example from Matt 7:26 above: the scribe realised only after he had written the first ouai that a paragraph break was needed. In the following example it is not clear how the scribe read the phrase en ekeinw tw kairw: Matt 11:24–25, folio 73.6 (NT 6), line 7.19–25 plhn legw umin anektoteron este gh sodomwn en hmera krisews h soi en ekinw tw kairw apokriqis o is+ eipen exomologoumai

Both line 23 (en ekinw tw kairw) and 24 (apokriqis o is+ eipen) are treated as new paragraphs and, indeed, both phrases function regularly as the natural start of a new section. Whether the scribe simply made a mistake repeating the extrusion for line 24 or whether he, on

104

Nomina Sacra, Ligatures, Itacisms, Text-divisions

second thoughts, took line 24 as the real start of the paragraph and line 23 as belonging to the previous sentence, is impossible to determine. There is one example where we find the opposite phenomenon to the examples above, namely the removal of a paragraph break. John 10:7, folio 79.7 (NT 54), line 5.8 autois . (v 7) eipen oun o is+ amhn

corrected to: autois . (v 7) eipen oun autois palin o is+ amhn

Here the scribe had started a new paragraph in the usual way with a half-open line and ekthesis of the new line, but, in a later correction by the scribe or a scriptorium-corrector, the first words are erased and rewitten on the previous, half empty previous line, the last letters in a small script. The corrector wants to include the words autois palin and removes both the empty space on the previous line and the ekthesis.87 The following three cases are all instances of a paragraph division in the wrong place: Matt 14:35, folio 74.1 (NT 9), line 3.14–19 nhsaret (v 35) kai epignontes auton oi andres tou topou apestilan eis olhn thn pericwron ekinhn kai prosh-

Matt 15:29, folio 74.1 (NT 9), line 6.35–37 kai metabas ekiqen o is+ hlqen para thn qalasan

Mark 13:3, folio 76.2 (NT 26), line 6.17–23 (v 3) kai kaqhmenou 87 For the reconstruction of the two stages I relied on Tischendorf‘s notes. The notebook by Milne and Skeat does not contain a comment on this passage, which may imply that they agree with Tischendorf.

Paragraphing

105

autou eis to oros twn elaiwn katenanti tou ierou ephrwta auton kat idian o petros kai iakwbos kai

The second and third example are preceded by a partially open line, and the latter even attracted the Eusebian section number which was added later by the scribes but was due at the start of the verse. As can readily be seen, the paragraph break occurs in the middle of a single sentence. Only in the first example does some sort of break within the sentence take place, as the verb apestilan is the first main verb after the participial construction. The last example also involves a textual variant: Luke 2:1–2 folio 76.6 (NT 30), line 3.26–35 egeneto de en tais hmerais ekinais exhlqen dogma para kaisaros agoustou apografesqe pasan thn oikoumenhn (v 2) authn apografhn egeneto prwth hgemoneuontos

The transcription of the text of the original scribe has been given above, and the genesis of the textual variant authn apografhn, a singular reading of Sinaiticus, is easily explainable. The words authn apografhn are assimilated to the previous words thn oikoumenhn, possibly helped by the presence of the article before apografh in the exemplar authhapografh.88 The subject of egeneto remains unexpressed (awkward, but not impossible) and starts off the paragraph “it happened first when...” This whole variant started either through a misunderstanding of the syntax in that auth (h) apografh was allowed to join itself to the previous syntactical unit, or, Westcott-Hort, Notes on select readings, s.l., explain the variant solely on the basis of mechanical assimilation to the preceding words and dismiss the possibility that an article before apografh played a role. 88

106

Nomina Sacra, Ligatures, Itacisms, Text-divisions

more likely, with a simple error of sight, taking the article of apografh as the final nu of an accusative auth. Additionally, egeneto is the type of word that can start a paragraph, as the first paragraph break in our example shows. The resultant nonsense reading was corrected, either immediately by the scribe or later by a scriptorium corrector, by removing the two nus at the end of authn apografhn, but the paragraph— wrongly placed on account of the preceding error—was left intact.

Scribe A in 4 Maccabees As seen above, 4 Maccabees contains a contrast between the number of paragraph division of scribes D and A. When the actual paragraph breaks are judged on their quality, it appears that two out of six paragraphs by scribe A start at an awkward place. In contrast to this, within the section written by scribe D we could find no unusual paragraphs. 4 Macc 10:1, folio 42.5 (OT 60), line 3.43–48 qias orghs dikas (v 1) kai toutou ton aoidimon tqanaton karterhsantos o tritos hgeto parakaloumenos pol-

Not unlike Mark 13:3 above, the new paragraph starts after the introductory genitive absolute. Apparently the force of o tritos (referring to the third of the seven brothers, who is about to suffer martyrdom) lead the scribe to commence a new paragraph. 4 Macc 11:1, folio 42.5 (OT 60), line 6.5–11 ektemneis (v 1) ws de kai outos tais basanois kataikisqeis enapeqanen . o pemptos parephdhsen legwn

Again, the prominent numeral, which forms the start of the main sentence, is made into the beginning of a new paragraph, leaving the preceding temporal clause on its own. The paragraph following this

Paragraphing

107

example, which starts off dealing with the sixth brother, correctly includes the introductory clause in the new paragraph.89

Scribe A in Romans From the above examples it becomes clear that certain phrases tend to be seen by the scribe as warranting the start of a new paragraph (numerals in 4 Maccabees, egeneto, kai, de). We find this particularly emphasised in scribe A’s copy of Romans. Let us consider, for example, the folio with the most paragraphs, folio 81.3 (NT 64). The text in this folio covers Rom 5:16–8:5; the first paragraph occurs at line 1.6, Rom 5:17. In total we count 61 paragraphs on this folio and one paragraphus, line 8.45. By comparison, NA27 has six large breaks and eight minor breaks in this section of 55 verses. Words at paragraph breaks, folio 81.3 ... gar 10 ei gar 2 ou gar 2 oida gar 1 oidamen gar 1 wsper gar 1 ote gar 2 ... de 5 nuni de 3 ean de 1 ei de 1 alla 3 outws kai 3 outw nun 1 ... oun 1 ara oun 1 ti oun 3 tina oun 1 mh oun 1 ara oun 2 ... ara 1 ouden ara nun 1 h 2 ina wsper 2 89

4 Macc 4:21, folio 42.2 (OT 57), line 7.31.

108

Nomina Sacra, Ligatures, Itacisms, Text-divisions Words at paragraph breaks, folio 81.3 wste kai 1 wste o men 1 oitines 1 pollw mallon 1 touto ginwskontes 1 oti ouk oidate oti 1 mhde 1 anqrwpinon legw 1 talaipwros egw 1 eucaristw 1

From this table it becomes clear that the scribe often used gar as a marker for the start of a new paragraph. This tendency is also confirmed by two of the three exceptions. The gar in Rom 7:1, line 5.3, does not receive a special treatment, but the other two do. In Rom 7:7, line 6.9, punctuation in the form of a raised dot is present (one out of only two on this folio), and the gar in Rom 8:5, line 8.45, receives a relatively rare paragraphus. For the particle de more exceptions exist. In this folio of Romans, paragraphs have taken over the function that is normally performed by punctuation: the paragraph breaks are frequent and at times break up longer sentences. Paragraphs do not function as a tool to give structure to the larger discourse; here they function at a much lower level. Concluding observations on the use of paragraphs Having surveyed the use of paragraphs in Sinaiticus, we can now answer the relevant question of whether the paragraphing in Sinaiticus is a testimony to the habit of the scribe or of a larger scribal tradition. The strongest indications that the paragraph breaks are the work of the individual scribes are the differences between two scribes within a single book. Four Maccabees shows a marked difference in both the quantity and quality of the paragraph divisions, and also the replacement leaves in the NT contain a different number of paragraphs from their immediate surroundings. Also, those paragraphs that start in dubious places within a sentence are better explained as occurring in the context of a scribe who makes the paragraphs up as he goes along rather than being the result of a scribal tradition.90 The Here the assumption is that a scribal tradition will produce a sensible and meaningful system of paragraphs. 90

Eusebian Apparatus

109

analysis of the quality of paragraph breaks demonstrates the influence of signal words. Also the breaks in the densely paragraphed folio of Romans affirms that signal words play a large role. The fluctuation in paragraphing within a single book, such as the gospel of Luke, is easier to explain by swings in the scribe’s attitude than by fluctuation within the tradition. However, the marked differences between books and groups of books are less easy to fit into this model. Why has Matthew, as a whole, a denser paragraphing than Luke, and the Pauline corpus than the Catholic Epistles? Also, the text of Revelation is not so unclear as to make a reasonable level of paragraphing impossible, yet the frequency of paragraphs is very low. It seems, therefore, that some influence from outside the scribe is present and that it is reasonable to assume that the exemplar, or exemplars, played a role.91 The current situation in Sinaiticus is best seen as the result of both the individual practices of the scribes and influence from the exemplar, though the first of these two factors seems to be more important. We found a considerable number of inexplicable paragraph divisions in the work of scribe B, while in scribe A’s work these were found less often. However, scribe A is certainly not consistent in his paragraphing and tends to fluctuate considerably from book to book. There is no evidence that these two scribes used paragraphs to structure the text on a discourse level. Frequently, paragraphs appear to have been inserted only because of the presence of certain signal words. Scribe D has few or no dubious paragraph divisions and overall does not use a dense paragraphing. Only in places where the text needs stretching out, do the number of paragraphs increase dramatically.

EUSEBIAN APPARATUS Introduction A feature of Sinaiticus that belongs to the earliest history of the codex is the insertion of the Eusebian apparatus, which was included, according to Milne and Skeat, by the original scribes themselves.92 As the sections are not present on the first replacement sheet in Matthew, though found on the In contrast to paragraphing in prose text, the division into stichoi in the poetic sections is very strongly influenced by tradition. Though the occasional combining of two stichoi into one does occur, the poetic lines of Sinaiticus follow the wider tradition closely. This warrants the conclusion that the line division must have been part of the copying process. 92 Milne-Skeat, 36. 91

110

Nomina Sacra, Ligatures, Itacisms, Text-divisions

folios in between the two replaced folios, it follows that these must have been included before the replacement occurred. They are present on the second replacement folio of Mark/Luke.93 Our codex contains most of this system, though the accompanying canon tables are not present (and possibly never were)94 and neither is the explanatory letter to Carpianus. Furthermore, in Luke, all the section numbers after 106 are missing except for a single one in a different but very early hand.95 The presence of the Eusebian apparatus has often functioned in attempts to date Sinaiticus; the manuscript cannot be older than the date at which Eusebius issued his innovation. To my knowledge, the apparatus itself has never been studied as a source of information on the earliest history of the manuscript. The logic behind the system is relatively easy to understand. In order to find a passage, or passages, in other Gospels that contain parallel material, the reader moves up in the current passage, finds the first combination of section and table number in the margin, and turns to the correct table. In this table, the related section numbers in the other Gospels are written on the same line as the number of the section which was found in the margin. Eusebius compiled ten tables, each containing different combinations of Gospels. One does need the tables in order to make sense of the section and table numbers.96 At this moment, any study of the Eusebian apparatus of a given manuscript will suffer from a very serious drawback, which is that hardly any comparative data from other manuscripts are available in a reliable and consistent presentation. I do not know of any systematic study of the variations within the Greek Eusebian apparatus apart from the study of Nestle and a single article by Nordenfalk.97 The Eusebian apparatus has not See also Chapter 2, page 41. See Chapter 1, page 32. 95 Luke 21:24, folio 78.4 (NT 44), line 8.46. The section number en§ is obviously wrong, the epsilon should have been a sigma. 96 An alternative and quicker system, found predominantly in Syriac but also in some Greek manuscripts, is to write the parallel sections in the lower margin of a page, e.g. for section 3/5 in Matthew, Mq g : Mr . : Lo b : Iw . , thus making the tables unnecessary. Codex Basiliensis (E) has the apparatus in the left margin, often with the table numbers, but also includes the comparative lines in the lower margin. The manuscript does not contain the tables. 97 Nestle’s study in three parts traces down the history of the Eusebian apparatus as printed in various editions of Greek, Latin, Coptic, and Syriac texts. He discusses a few problem spots and informs the reader that the section division found in the sixth edition of Nestle’s text contains his reconstruction of Eusebius’s original system, based on a comparison of ℵ A C D b f g got syr copt. E. Nestle, 93 94

Eusebian Apparatus

111

been part of Legg’s volumes on Matthew and Mark and has also been left 98 out in the IGNTP volumes on Luke. Nestle-Aland’s 27th edition states that its version of the Eusebian apparatus is identical to “the standard critical edition by Eberhard Nestle in the 25th and 26th editions,”99 but differs in some of its details from the apparatus found in the Stuttgart Vulgate.100 The only readily available source for the section and table numbers as found in Greek manuscripts and their actual placement are the volumes by Reuben Swanson, which are not always free from error.101 However, as it is the only available tool, we rely on the information as given by Swanson.

“Die Eusebianische Evangelien-Synopse”, Neue kirchliche Zeitschrift 19 (1908): 40–51, 93–114, 219–32. He also provides examples of scribal errors made when inserting the apparatus. Nordenfalk studied canon tables from the perspective of the history of art; C. Nordenfalk, Die spätantiken Kanontafeln (Goteborg: O. Isacsons boktryckeri a.-b., 1938). Later he discussed also some textual problems in the tables “The Eusebian Canon-Tables: Some Textual Problems”, Journal of Theological Studies 35 (1984): 96–104. 98 S.C.E. Legg, Euangelium secundum Marcum (Oxonii: E Typographeo Clarendoniano, 1935), and Euangelium secundum Matthaeum (Oxonii: E Typographeo Clarendoniano, 1940). International Greek New Testament Project, The Gospel according to St. Luke, 2 vols. (Oxford: Clarendon Press, 1983–87). 99 NA27, introduction 79* n. 10. The reference to Nestle suggests that the ‘standard edition’ is the one described in Nestle’s study of 1908. As Nestle reconstructed Eusebius’ sections using some Greek uncials and versional evidence, there is still no edition describing the transmission of the sections in the Greek tradition. 100 As printed in B. Fischer et al., Biblia Sacra iuxta Vulgatam Versionem, 4th ed. (Stuttgart: Deutsche Bibelgesellschaft, 1994). The preface to the first edition suggests that the Eusebian apparatus used in the edition are freshly collated from the Latin manuscripts: ‘In the Gospels the numbers of the Eusebian Sections and Canons, omitted by Clement, have been inserted in accordance with the manuscripts’ (Prefaces, xxx). 101 Each Gospel has its own volume: R.J. Swanson, New Testament Greek Manuscripts. Variant Readings Arranged in Horizontal Lines against Codex Vaticanus (Sheffield: Sheffield Academic Press, 1995). An Errata list is published on the Web (http://www-user.uni-bremen.de/~wie/texte/Swanson-errata.html), which, alas, has become too long to be practical.

112

Nomina Sacra, Ligatures, Itacisms, Text-divisions

The positioning of the Eusebian apparatus in the margin

Scribal customs and clear scribal errors The Eusebian apparatus is always positioned in the left margin before the line with which, or within which, the section is supposed to start. In many cases the scribe marks the start of the section in the line with two dots, which leaves no doubt as to where the section is supposed to start. Very often the start of a section coincides with the place where a scribe commences a new paragraph, which is, of course, a natural place for the Eusebian apparatus. When a paragraph break occurs shortly after the start of the Eusebian section, the paragraph break may attract the apparatus as in the following examples: Mark 10:21, folio 75.8 (NT 24), line 7.25–31

rh/b

o de is+ embleyas autw hgaphsen auton kai eipen autw eti en se usteri upage osa eceis pwlhson kai dos tois ptwcois kai

The section starts in the rest of the textual tradition with the words eti en se usterei; the paragraph break of Sinaiticus is slightly artificial but nevertheless attracts the Eusebian apparatus. Mark 13:3, folio 76.2 (NT 26), line 6.17–24 kai kaqhmenou autou eis to oros tw( elaiwn katenanti tou ierou ephrwta auton rlh/b katidian o petros kai iakwbos kai iwannhs kai an-

The paragraph break occurs in the middle of a syntactical unit, but the section number is placed before the start of this paragraph, even though the section starts in the remainder of the tradition with the first quoted line kai kaqhmenou.

Eusebian Apparatus

113

Many times the section number does not coincide with the start of a paragraph. In such cases the Eusebian apparatus is placed immediately before the line where the section starts. However, if this line contains only a single word or an introductory formula, then the apparatus is often placed before the second line of the section. Luke 6:31–32, folio 77.1 (NT 33), line 6.24–31 nd/e

ne/e

mh apaitei . kai kaqws qelete ina poiwsin umin oi anqrwpoi kai umis poieitai autois omoiws . kai ei agapate tous agapwntas umas

On line 24, the scribe places the Eusebian apparatus before the line where section 44 starts, but the numbers for section 45 are placed a line below the actual start, on line 30. In both places, the only word on the line that belongs to the new section is kai. Another reason why the Eusebian apparatus is sometimes not placed accurately is that there is simply not enough space at the original location. At John 6:49, the section number is driven one line lower by the previous numbers; there are three Eusebian sections here within the space of six lines.102 Similarly, the original space can be occupied by a correction, such as at Matt 10:17.103 Another type of error can occur when a section is written too far down: Section 122 in John 13:23 and 223 in John 21:12 are both written in the place of the following section, thus skipping a whole section without increasing the section number.104 In the first case, the next section number (123) is subsequently left out in order to get the numbering back on track, though the table number for section 123 is used with section 122. In John 21, section 224 is given a unique position between its proper place and the next section. The reverse happens in Mark 9:33, where the section and canon number of the next section 95/2 are placed at the position of John 6:49, folio 79.4 (NT 51), line 8.7. Matt 10:17, folio 73.6 (NT 6), line 1.1. On the same page the Eusebian apparatus is added to a correction in the lower margin. 104 John 13:23, folio 80.2 (NT 57), line 1.14. John 21:12, folio 80.6 (NT 61), line 2.32. 102 103

114

Nomina Sacra, Ligatures, Itacisms, Text-divisions

94/10.105 Once, at John 6:38, the scribe repeats the section number which he used already at 6:35 instead of increasing the section by one. He writes again 56 instead of 57 whilst adding the correct table for section 57.106 At Mark 10:41 a related phenomenon is seen. Originally section number 113 was written, corrected immediately to the correct 114. Section 113 is missing in the manuscript and it appears that the scribe was apparently expecting to write section 113, but realised only after more careful reading of his exemplar that he had arrived at 114. This suggests that the problem of the missing section 113 was already present in the exemplar. A difficulty arises on the first page of Matthew: at Matthew 1:18, it is hard to determine where the scribe intended the Eusebian section to commence. Matt 1:18, folio 73.1 (NT 1), line 3.11–20 g/e

d/i

tou de iu+ cu+ h genesis outws hn mnhsteuqishs ths mhtros autou marias tw iwshf prin h sunelqin autous eureqh en gastri ecousa ek pn+s agiou Ú (v 19) iwshf de o anhr auths di-

According to the placement of the apparatus, one would expect that the reference 4/10 was to mnhsteuqisths which can indeed be taken as the beginning of a section. However, the section starts in most of the remainder of the tradition at verse 19 and indeed we find here the characteristic double dot that so often serves to indicate the start of a section. It may well be that the scribe who added the apparatus realised his initial misplacement and corrected it by adding the double dot at the correct spot six lines lower. A displacement error can also happen when the apparatus is placed at the wrong column but roughly on the correct height within this column. This happens with Mark section 37, placed before the third column instead

105 106

Mark 9:33, folio 75.8 (NT 24), line 3.4. John 6:38, folio 79.4 (NT 51), line 6.42.

115

Eusebian Apparatus

of the fourth,107 and probably section 72, which is also misplaced by roughly one column.108

Substantial deviations in placement In order to get an impression of the general accuracy of the placement of the Eusebian apparatus, we have compared each position in Sinaiticus with that in NA27 and, if a discrepancy existed, with the data as given in the Swanson volumes. Excluded are all the instances where the section and table numbers are displaced by only a single line, but which could still be construed to refer to the same starting point as in NA27. The full results are given in Appendix I: “List of all the differences in position in the Eusebian apparatus between Sinaiticus and NA27.” Matt (355 sections) Total displacements 26110 Singular 7 Near singular (1-3 mss) 6 With other manuscripts (4- 4 7 mss) With many other 7 manuscripts Missing apparatus114 - 115

Mark (233)

Luke (106 [342])109

John (232)

30111 12 8 5

3112 2 1

18113 5 6 6

3

-

1

11

1

4

Mark 4:2, folio 75.4 (NT 20), line 3.1. Mark 7:6, folio 75.6 (NT 22), line 5.34. 109 Traditionally Luke has 342 sections, all the section after 106 are missing except section 256/10 before Luke 21:24, folio 78.4 (NT 44), line 8.46. 110 At two places in Matthew, NA27 has a slightly different placement from all of the Greek tradition as well as from the Vulgate: Matt 13:57 (142/1), Matt 24:16 (248/2). These two cases have not been included in the table. 111 At two places in Mark where Sinaiticus and NA27 differed in their positioning of the apparatus, Swanson provides no relevant data (Mark 10:17 107/2 and 10:46 116/2). 112 At one place in Luke, NA27 has a different placement from all of the tradition: Luke 6:36, 37 (56/2) which is not included in the table. 113 At three places in John, NA27 has a slightly different placement from all of the Greek tradition, in the second it has support from the Vulgate: John 12:27 (132/10), John 14:25 (132/10), John 19:25 (202/10). These three cases have not been included in the table. 114 See Appendix II ‘Places where the Eusebian apparatus is missing’. 107 108

116

Nomina Sacra, Ligatures, Itacisms, Text-divisions

Manuscripts occurring more than once in the “near singular” and “with other manuscripts” category: Matthew: Q 124 788 (4 times), D M Y (twice). Mark: Y (5 times), Q 579 (4 times), G G 1424 (3 times), 2 118 (twice). Luke116: Only one deviating position, shared with E H K S Q 565 579 1346 John: 1071 (4 times), G K L S Y f13 2 (3 times), E H U L Y W 28 157 788 1346 (twice)117 Deviating section and table numbers Another area of potential confusion is the combination of section with table number. In order for the Eusebian system to work, a section must be combined with the correct table. It is possible that in the course of transmission various varieties of these section/table combinations came into existence, possibly—but not necessarily—in combination with adjusted tables as well. The full list of deviating section numbers is given in Appendix III: “Deviating table numbers in the Eusebian apparatus.”

Total deviating table numbers Singular Near singular (1–3 mss.) With other manuscripts (4-7 mss.) With many other manuscripts Missing table number Deviating section numbers

Matt (355 sections) 19118

Mark (233)

Luke (106)

John (232)

4

5

18

7 10 3

2 1

2 3 -

4119 10 4

1

-

-

-

-

1 -

-

4 3120

115 Except the 18 places omissions on the first replacement folio and the 17 omissions on the second. 116 Provided that Swanson is interpreted correctly here. 117 The occurrence of 788 and f13 is due to Swanson’s way of listing the evidence. 118 The specification of these 19 table numbers add up to 21 as the scribe made on two places a correction to the canon number. Both the first and the second version are included in the totals. 119 For John 6:30 53/1, Swanson provides no data and is assumed here a SR. 120 The three are John 6:38 56/1, John 13:23 122/10 (see above page 113), and John 19:21, where the scribe wrote si without a table number instead of s/i.

Eusebian Apparatus

117

Manuscripts occurring more than once in the “near singular” and “with other manuscripts” category: Matthew: E (6 times),121 124 (4 times), S W (3 times), M Y 788 (twice). Mark: One deviating table number shared with other manuscripts: E S U W 1071. Luke: Three deviating table numbers shared with other manuscripts: A (twice), L 788 (once). John: N 788 (6 times), 124 (5 times), G 118 (3 times), E L P W 28 (twice). Discussion of deviating placements and deviating numbers The totals for each Gospel reveal rather striking differences between the four Gospels. Matthew stands out from the other three Gospels in that not a single canon or section number is missing, and also two corrections are made on the apparatus.122 Elsewhere this only happens once (in Mark).123 On the other hand, the book contains a fair number of deviating table numbers. All of these are “possible” tables, that is, no referral is made to a table in which Matthew does not function (tables 8 and 9). An interesting situation arises at Matt 16 with regard to sections 162– 164 (sections 165–182 are absent in the following replacement folio). In the majority of the tradition, canon 162/5 covers Matt 16:2–3 and 163/6 Matt 16:4. However, a substantial part of the textual tradition does not read anything after the words o de apokriqeis eipen autois in verse 2 and omits verse 3 completely, which leaves only the introductory formula to section 162. As was to be expected, this situation gave rise to considerable confusion, and different groups of manuscripts follow different traditions. If we consider sections 163 and 164 together, Sinaiticus is only joined in its version of the placement problem by Codex Bezae (D) and Codex Campianus (M), though more manuscripts join in at section 164. Interestingly, the scribe makes a correction to the table number of section Codex Basiliensis (E) seems to have swapped tables 5 and 6 in the apparatus present in the left margin. In the lower margin this manuscript spells out the related section numbers for each other gospel, often affirming the traditional tables. However, two of the six agreements are found in the written-out apparatus in the lower margin. 122 The exceptions are, of course, the missing numbers on the two replacement leaves in Matthew (folio 74.2 and 74.7) which in their present form do not contain the Eusebian apparatus. At Matt 15:24, 158/6 (shared with E) was changed to the majority reading 158/5, for the other correction see the main text. 123 Mark 11:41, section number 113 corrected to 114, while 113 remains missing. 121

118

Nomina Sacra, Ligatures, Itacisms, Text-divisions

163, which was initially 5 or 6 (probably the latter), but changed to 2. Canon 163/6 is the combination found in most manuscripts that have section 163 at an earlier place. The correction to table 2 is indeed correct for the section where Sinaiticus places section 163 (164 in most other manuscripts). The change of table number seems to reflect a conflation between two traditions of the apparatus. The confusion suggests that the Eusebian apparatus of Sinaiticus is taken from a manuscript that included verses 2b–3. An area for further research might be the relation between the Eusebian apparatus of Sinaiticus with minuscules 124 and 788, both members of family 13. Both in the placement of the apparatus as well as in deviating canon numbers, the three manuscripts share readings. However, only after a complete comparison of the apparatus more can be said about any shared tradition.124 Mark contains relatively few deviating or missing canon numbers. The main problems in Mark are the many places at which the apparatus is missing altogether (though not on the replacement folio of this Gospel) and the numerous places where one finds deviating positions, 12 of these unique to Sinaiticus. The textual confusion in Mark 9:43–48 is considerable, with verses 44 and 46 absent in Sinaiticus, as in many witnesses. Also the placement of sections 100 and 101 shows considerable variation. Sinaiticus has section 100 before the start of verse 45 and skips section 101. It is not unlikely that these two consecutive irregularities both arose because of the textual confusion.125 In the wider tradition, Mark 15:28 has its own section and table number, 216/8. This verse, however, is probably an intrusion from Luke and is absent in a number of witnesses, including Sinaiticus. Yet, the section number of the missing verse is present and takes the position of the next section, 217, which is subsequently left out. The correct way of 124 The method to be followed in order to depict the real relationship between the apparatuses of these manuscripts is that of a full comparison of the Eusebian apparatus, not limited to agreement in error only, as we have done above. The theoretical framework for this is no different from that for establishing relations between actual texts. See E.C. Colwell, Studies in Methodology in Textual Criticism of the New Testament (Grand Rapids: Eerdmans, 1969), chapter 1, 1–25, originally published as “Significance of Grouping of New Testament Manuscripts”, New Testament Studies 4 (1958): 73–92. Also Gordon Fee’s remarks in “Codex Sinaiticus in the Gospel of John: A Contribution to Methodology in Establishing Textual Relationships” in E.J. Epp and G.D. Fee, Studies in the Theory and Method of New Testament Textual Criticism (Grand Rapids, Mi.: Eerdmans, 1993), 223–27 (originally published in New Testament Studies 15 [1968/69]). 125 Nestle (“Die Eusebianische Evangelien-Synopse”, 47–48) dismisses the witness of Sinaiticus to the confusion in these verses too easily.

Eusebian Apparatus

119

numbering would have been to leave 216/8 out and continue with 217/6.126 The most likely way for this situation could have arisen is that the Eusebian apparatus was copied from a manuscript that included Mark 15:28 into a manuscript that did not include it. This may have happened when the apparatus was copied into Sinaiticus, but it could also have happened in the ancestry of the exemplar which provided the Eusebian sections. The Gospel of John shows a high number of table numbers that are not widely attested, though the number of agreements of Sinaiticus with minuscules 124 and 788 and Codex Purpureus Petropolitanus (N) is remarkable. This should function as a caution against ascribing all the deviations to the scribe and allow for genuine variation within the tradition of the Eusebian apparatus. However, only minuscule 788 shows up in the list of manuscripts supporting deviating positions for the section numbers. Four section and canon numbers are missing altogether, less than in Mark but relatively more than in Matthew and Luke. In one instance, the scribe writes the canon number after the section number, creating the number 210 instead of 200/10. In another instance, just as in Matthew, a section number is repeated with a table number corresponding to the correct rather than to the repeated section number. The distribution of the various phenomena over the four Gospels is very uneven. Matthew has all the section and canon numbers, Mark misses 11, Luke misses one before 106 and all the numbers after this point, and John misses four. In John, we find three section numbers without a table number, elsewhere this occurs only once in Mark. In Matthew, seven times we have a unique combination of section and table numbers, considerably more than in Mark and Luke, but, making allowance for the lower number of sections, comparable with John. Some of the table numbers seem to be purely scribal errors, as they refer to a table in which the respective Gospel does not function. These are Luke 3:21, 13/4; John 13:16, 118/6; John 18:12, 162/5; John 21:1, 219/5. Remarkably, at two of these places Sinaiticus has support from one or more other manuscripts. Concluding observations on the Eusebian apparatus Scribes and Correctors states that the first 52 numbers of Matthew are by scribe A and were later retraced by scribe D who also adds all the other numbers. This may be correct, but it does not show up as a changed level

126

Nestle, ibid. 102.

120

Nomina Sacra, Ligatures, Itacisms, Text-divisions

of quality in the Eusebian apparatus.127 All four Gospels share some errors, in that there is a percentage of unique table numbers and/or positioning. However, they also show different patterns in some of the types of errors. Though it is possible that a scribe was precise in Matthew, not skipping any section, and much less precise in Mark, an answer may also be sought in the exemplar or exemplars that provided the Eusebian apparatus for Sinaiticus. This means that, as far as the Eusebian apparatus is concerned, not all the four Gospels share the same tradition history. What this tradition history looked like is difficult to determine. It may be that the Eusebian apparatus was copied into the exemplar by different scribes or that the Eusebian apparatus was copied into Sinaiticus from different exemplars. These exemplars may have been different four-gospel books or even individual gospel manuscripts.128 From our survey of the Eusebian apparatus in Sinaiticus, it follows that this version of the Eusebian apparatus is probably not an exact and faithful representation of the original system.129 The handling of passages such as Matt 16:2–6 and Mark 15:28 betrays a corruption in the transmission history of the apparatus between its inception and inclusion in Sinaiticus.

NUMBERING SYSTEMS AND KEPHALIA In three different books a section system is used that must have been added by the scribes or a contemporary hand. In Ecclesiastes one finds the numbers a to d in the margin in the first two folios of the book, after which the numbering stops at Eccl 4:9. In the next book, Song of Songs, the numbers a to d are also found, but here the last number can easily be construed as covering the remainder of the book. The numbering in these two books is in the same bookhand as the main text of the two books. In Though I had genuine doubts whether Milne and Skeat were correct in their judgement on the retracement, it clearly proves to be the case when the apparatus is checked in the original. That a change of hand is also very likely is also indicated by the sudden absence of paragraphi after section 52. It may be, though, that not all the Eusebian apparatus after this section are by scribe D, at some places I suspect that scribe A took over again. The scribes were not totally consistent in the bookhand that they used; see for example the different alphas before column 4 of folio 74.8 (NT 16). 128 To my knowledge there is no evidence of a manuscript containing only one or two Gospels and the Eusebian apparatus, which makes the last option very unlikely. 129 Admittedly, the critical work of establishing the ‘original’ Eusebian system or mapping its transmission history has not been attempted in this study. 127

Numbering Systems and Kephalia

121

the NT, Acts is numbered a to mb (=42), and, as in the OT, this system is also discontinued (last number at Acts 15:40). Here the numbering is in a smaller hand, which was ascribed by Milne and Skeat to scribe A. It has been noted that in the numbering for Acts, Sinaiticus follows closely the secondary numbering found in Vaticanus.130 Vaticanus contains two almost unique numbering systems, one of these in a contemporary hand,131 the second in a hand which is unlikely to be earlier than the sixth century and which does not cover each book.132 In Acts, this younger system divides Acts into 69 sections. Westcott and Hort linked this system of divisions with that found in the Latin Codex Amiatinus, and argued on this basis that the origin of both Vaticanus and Sinaiticus should be sought in the West.133 The section numbering in Acts Not all the numbers are present in Sinaiticus; missing are a, g–z, and all numbers after mb. The sections in Sinaiticus are exactly paralleled by Vaticanus in 31 out of 36 places. Vaticanus skips Sinaiticus’s section m (40) and therefore has a section number that is one count lower for the sections ma and mb, for which places there is agreement in position but not in number. Wordsworth and White give in their edition of the Latin Bible the capitula for several of the important Latin manuscripts, but they do not indicate from where to where a particular section runs.134 The total number of section headings of Codex Amiatinus is 70, which is very close to the 130 See e.g. C.R. Gregory, Textkritik des Neuen Testamentes, vol. 1 (Leipzig: J.C. Hinrichs, 1900), 33. 131 The same numbering system has also been noticed in the margin of Codex Zacynthius (X) for the gospel of Luke and in minuscule 579. 132 In some books, as in Jeremiah, the secondary numbering simply enlarges the first numbering. I do not know whether this is because the scribe responsible for the secondary numbering had no secondary numbering available for this book or because the two system were identical in these books. 133 Westcott-Hort, 266. Berger (who fails to notice the later date of the secondary system in Vaticanus) supposes that the system came into Latin from the Greek. Histoire de la Vulgate pendant les Premiers Siècles du Moyen Âge (Paris: Hachette et cie, 1893), 313. 134 J. Wordsworth et al., Nouum Testamentum Domini Nostri Iesu Christi Latine secundum Editionem Sancti Hieronymi, Part Three: Acts - Catholic Epistles - Apocalyps of John (Oxonii: E Typographeo Clarendoniano, 1954). Not seen: Donatien de Bruyne, Sommaires, Divisions et Rubriques de la Bible latine (Namur: A. Godenne, 1914).

122

Nomina Sacra, Ligatures, Itacisms, Text-divisions

number of 69 found in Vaticanus, especially bearing in mind that Vaticanus has possibly skipped section 40. Another group of Vulgate manuscripts has a section division that runs up to 74, of which the first 41 sections are exactly the same as in Amiatinus. On the basis of the wording of the capitula, which often contain the first words of a section, it is possible to check whether the Latin and Greek sections coincide. Out of the 36 sections in Sinaiticus, 28 times agreement in number and position exists, the eight discrepancies ranging from a difference of half a verse to a couple of verses. The agreement in numbering is not distorted. It is justified, therefore, to conclude that basically the Greek system found in Sinaiticus and the younger hand of Vaticanus is similar to the Latin system.135

Kephalia in Acts Besides the discontinued section numbering, a series of titles is also found in the upper margin in Sinaiticus, assigned to scribe A by Milne and Skeat.136 These titles cover the whole book and do not show any direct relation to the section numbering. The Greek texts cited in the following list refer to the probable start of the kephalion. In the text itself no marking can be found which indicates the actual start. 1. ta peri ton petron  iwannhn  tou ek koilias cwlou folio 85.8 (NT 101r), col 3; Acts 3:1 2. ta peri ton anannian  thn gunaika sapfeiras folio 86.1 (NT 102r), col 4; Acts 5:1 3. ta peri ton stefanon folio 86.2 (NT 103r), col 4; Acts 6:9 Von Soden’s treatment of the numbering systems is wholly inadequate. He states, for example, that the earliest section numbering of the gospels found in Vaticanus, is also seen in the paragraph divisions of Sinaiticus, which is not true. Likewise, he assumes that the numbering in Acts goes back to a combination of the kephalia and the upodiaireseis. H. von Soden, Die Schriften des neuen Testaments: in ihrer ältesten erreichbaren Textgestalt I. Teil: Untersuchungen; I. Abteilung: Die Textzeugen (Göttingen: Vandenhoeck und Ruprecht, 1911), 440–48. It seems to me that the systems in the two uncials are completely independent from the Euthalian system of kephalia (in 36 or 40 chapters) and its subdivision into upodiaireseis. The similarities between the two systems are more likely due to the inherent narrative structure of the book of Acts. 136 Milne-Skeat, 38. 135

Numbering Systems and Kephalia 4. ta peri simwnos tou magou folio 86.3 (NT 104v), col 6; Acts 8:9 5. ta peri filippon folio 86.4 (NT 105r), col 1; Acts 8:26 6. ta peri saulon folio 86.4 (NT 105v), col 6; Acts 9:22 (first verse of column) 7. ta peri ainean folio 86.4 (NT 105v), col 7; Acts 9:32 8. ta peri tabiqan folio 86.4 (NT 105v), col 8; Acts 9:36 9. ta peri kornhlion folio 86.5 (NT 106r), col 1; Acts 10:1 10. ta peri barihsou tou magou folio 86.6 (NT 107v), col 8; Acts 13:6 11. ta peri ludias folio 87.1 (NT 110r), col 2; Acts 16:13 12. ta peri tas aqhnas folio 87.1 (NT 110v), col 8; Acts 17:16 13. apo twn aqhnwn hlqen o paulos eis korinqon folio 87.2 (NT 111r), col 3; Acts 18:1 14. ta peri ton apellhn folio 87.2 (NT 111v), col 6; Acts 18:24 15. ta peri twn exorkistwn folio 87.3 (NT 112r), col 1; Acts 19:13 16. ta peri demhtrion ton argurokopon folio 87.3 (NT 112r), col 3; Acts 19:24 (v23 previous column) 17. ta peri eutucon ote epesen apo tou tristegou folio 87.3 (NT 112v), col 6; Acts 20:7 18. tois presbuterois ths asias paulou diamarturia folio 87.3 (NT 112v), col 8; Acts 20:17 19. ta peri paulon ote anebainen is ierosoluma folio 87.4 (NT 113v), col 5; Acts 21:15

123

124

Nomina Sacra, Ligatures, Itacisms, Text-divisions 20. ta peri twn agnisqentwn folio 87.4 (NT 113v), col 7; Acts 21:26 21. paulou apologia pros ton laon epitreyantos tou ciliarcou folio 87.5 (NT 114r), col 1; Acts 21:39 22. tou paulou epi thn epaurion dialektos pros ton ciliarcon peri ths apodhmias ths eis ton arcierea  ton sunedreion folio 87.5 (NT 114v), col 5; Acts 22:6 23. epistolh klaudiou tou ciarcou pros fhlika ton hgemona folio 87.6 (NT 115r), col 2; Acts 23:26 24. tou paulou apologia pros fhlika ton hgemona folio 87.6 (NT 115r), col 4; Acts 24:10 25. tou paulou h pros ton fhston apologia kai ekklhtos folio 87.6 (NT 115v), col 8; Acts 25:9 26. tou paulou apologia pros agrippan ton basilea folio 87.7 (NT 116r), col 3; Acts 26:1 27. ta peri tou ploun tou paulou anercomenou eis rwmhn folio 87.7 (NT 116v), col 8; Acts 27:1

The particular format of these titles is, to my knowledge, unique to Sinaiticus. Thirteen of the titles have the format ta peri + accusative, while five others read ta peri + genitive. In the traditional kephalia to the Gospels, the format is always peri + genitive, and also in the wider Greek literature no titles in the particular format of Sinaiticus could be found. These kephalia are unique in their actual wording, and seem also unrelated to the wording of the Latin capitula. Because of this uniqueness, it cannot be excluded that these titles were an invention by someone within the scriptorium. The section numbering in Ecclesiastes and Song of Songs Since a similar section numbering was inserted by an original hand in Sinaiticus and a later hand in Vaticanus in the book of Acts, it is worthwhile

125

Numbering Systems and Kephalia

to see if the same is also true for the numberings of the other two books with numbering in Sinaiticus: Ecclesiastes and Song of Songs. The first four numbers of Ecclesiastes in Sinaiticus and Vaticanus are indeed positioned at exactly the same places. After this, Vaticanus continues with three further sections which do not have counterparts in Sinaiticus. However, in Sinaiticus a series of paragraphi can be found and one might wonder whether these could have something to do with a division into sections. From the table below it is clear that neither the paragraphi nor the Latin system used in Amiatinus has any relation to the numbering system employed most fully in Vaticanus. Ecclesiastes 1:1 1:16 2:2 2:18 3:1 3:18 4:9 4:13 5:18 6:9 7:2 7:3 7:11 7:29 8:8e 8:14 9:2b 9:11 9:15 9:18b 10:8 10:16 11:8e

ℵ a

B a

b

b

g

g

d

d

Amiatinus137 I II III IIII V VI

paragraphus VII e

VIII VIIII paragraphus paragraphus

X §

paragraphus XI paragraphus paragraphus z

paragraphus

In Song of Songs, the first four numbers in Sinaiticus and Vaticanus are also positioned at the same place in the text. The only difference is that Vaticanus has one extra section number, just before the end of the book. Capitula as given in the large Vulgate, vol. 11. Libri Salomonis id est Proverbia Ecclesiastes Canticum Canticorum, vol. XI (Rome: Typis polyglottis Vaticanis, 1957). 137

126

Nomina Sacra, Ligatures, Itacisms, Text-divisions

The large Vulgate edition does not record any capitula for this book. A comparison with Latin manuscripts is therefore not possible. Song of Songs 1:1 1:15 3:6 6:6 8:11

ℵ a b g d

B a b g d e

Concluding observations on the numbering systems and kephalia Sinaiticus has, in three different books, the section numbering which was added by a later hand to Vaticanus. It seems that in Sinaiticus we have only fragments of a system which actually covers large parts of the whole Greek Bible.138 The origin of this system must go back well into the fourth century or earlier. As the system also appears in Vulgate manuscripts of Acts, the question of the origin becomes even more complicated. Why is it only in Acts that the systems coincide? Unless new evidence is brought forward, it is impossible to tell whether the Latin system derives from the Greek or vice versa. The only certain thing is that the two traditions must have been in contact with one another. The two types of the Latin system are evidence of internal development within its transmission, for which a certain amount of time is needed.139 A possible point of contact could be, of course, Jerome’s translation of the Vulgate in the late fourth century. However, Jerome is silent on this issue, while he is, in contrast, clear on his borrowing of the Eusebian apparatus.140 Although none of the three books in Sinaiticus that contain part of the system have all the numbers as we find them later inserted in Vaticanus, they look more like a reflection of the full version of the system than individual elements that would later develop into a more comprehensive system. In the great majority of books where the system is present in Vaticanus, it is absent in Sinaiticus, but there is no evidence to suggest that in the days when Sinaiticus was written the numbering system only covered these three books. Most likely it was already completely developed, but incompletely included.

Though note the caveat mentioned in footnote 132. The oldest evidence recorded for the Latin system in Wordsworth and White is Codex Fuldensis (F) from the sixth century. 140 See Jerome’s introduction to his translation of the Gospels. 138 139

Conclusions

127

A common feature of all three discussed numbering systems is that they were added by scribe A in text written by himself. None of the three systems seems to have been included complete.

CONCLUSIONS The wide range of studies in this chapter does confirm the general impression of the attitude of each of the three scribes that one gains by reading Scribes and Correctors. At the same time we have filled in a lot of details and have extended the scribal profile to areas of scribal behaviour which were not dealt with by Milne and Skeat. The individual characteristics of the three scribes as they appear in this chapter can be summarised as follows. With regard to nomina sacra, scribe A shows a preference for the wider implementation of these contractions. He consistently uses the nomen sacrum for kurios, irrespective of whether it is in a sacral or profane use, and has examples of all the words that could be written as a nomen sacrum. In a large group of words, he wavers between the plene and contracted words and he seems to change his policy from book to book. The paragraphing of scribe A is inconsistent and not always correct. There are some corrections to paragraphing, indicating that the scribe was not completely indifferent to this aspect of the text. There is no real evidence that the scribe used nomina sacra or paragraph breaks to interpret the text. Nomina sacra are used irrespective of the sacral status of their referent, and paragraph breaks do not follow a definable pattern at all and may often have been inserted because of elements at the syntactical level rather on the basis of features at the discourse level of text. Scribe A has relatively few ligatures and is certainly not a strong speller. In Matthew, the profile of his orthographic errors changes dramatically. The change in spelling in Matthew is interesting in the light of scribal behaviour. Do alternative explanations exist for this change? If we also consider that the pattern of nomina sacra is slightly different for this book and that the superscriptions for Matthew follow a pattern not seen in the remaining Gospels, one starts to wonder if we are not dealing with a different and as yet unrecognised scribe. We could add two further arguments in favour of a fourth scribe. Firstly, the so-called B corrections occur predominantly in the Gospel of Matthew and could therefore be allocated to this fourth scribe. Milne and Skeat ascribe all these corrections to scribe A, but acknowledge that they form a distinct group of corrections. However, if we apply their criteria for distinguishing scribes to Matthew, we find a much stronger argument. In Scribes and Correctors, Milne and Skeat

128

Nomina Sacra, Ligatures, Itacisms, Text-divisions

point out that the coronis a scribe uses at the subscription of a book “amounts to his signature.”141 Interestingly, the coronis at the end of Matthew shows indeed a pattern that is fundamentally different when compared with scribe A’s normal design and does not appear at any other place in the manuscript. A good case can be made for a fourth scribe who is responsible for the Gospel of Matthew only and whose bookhand is virtually identical with that of scribe A. This scribe would be only distinguishable from scribe A by means of an analysis of his orthography and by the coronis. However, none of the palaeographic authorities has so far ever proposed a different scribe for Matthew, and it may well be that an explanation for the differences between Matthew and other work by scribe A must be sought in some outside influence. Possible factors may be a difference in the method of copying, influence from the exemplar, or inconsistency in scribal behaviour.142 The new investigation of the manuscript undertaken by the British Library will hopefully solve the problem of the scribe of Matthew, but at this moment one cannot state that the existence of a new scribe has been demonstrated beyond any doubt. Scribe B received a damning appraisal from Milne and Skeat and this is confirmed by our findings. His spelling is not good at all and also in the use of nomina sacra he is inconsistent in some of the forms (especially in the variations used for ierousalhm). The paragraphs in the first part of Isaiah do not betray any clear policy and at times run counter to the flow of the text. He makes free use of all types of ligatures and is the only scribe who uses a ligature for pro(s). In the previous chapter, we observed his consistency in the addition of the tituli. From the data reviewed in this chapter, however, it appears that this scribe is far from consistent in most of his practices. Finally, the work of scribe D shows some interplay between his understanding of the text and some of the scribal practices discussed in this chapter. He has no awkward paragraphs in the parts of the text we investigated and in the nomina sacra for kurios he shows an ongoing awareness of the nature of the referent. He is the most conservative in the employment of nomina sacra, though occasionally he does use a contraction for words such as dauid, ierousalhm, pathr, mhthr, and swthr. Ligatures are used freely in his section of Psalms, a part of the Milne-Skeat, 27. See also above, page 13. One could speculate that if scribe A copied an exemplar of Matthew written by scribe B, the result would be an orthographic pattern which improved on scribe B’s normal practice but would still differ from that of scribe A. The other differences are not explained by this guess. 141 142

Conclusions

129

codex that also receives specific ornamental treatment in the form of using red ink for the superscriptions to the individual psalms. The individual studies in this chapter give ample evidence for the thesis that it is possible to talk about distinctive, individual scribal behaviour. But to what extent does this chapter help us in understanding the manuscript as a whole and the way it was produced? First, it appears that the scribes possess considerable freedom in the execution of their task. Though they all use the same bookhand and adhere to the same general layout of a page, in many of the finer details they reflect their own scribal preferences. External control on the uniformity of some aspects of scribal practice is lacking. And, especially in the case of scribe B, this leads to a relatively weak result when it comes to the final form of the text. Secondly, inconsistency between different parts of the codex does not automatically mean that we are dealing with a different scribe; each scribe displays some level of variation within his work. This internal inconsistency cannot be fully explained by the influence of various exemplars either, although sometimes the exemplar may play a role. Thirdly, the last two studies in this chapter, on the Eusebian apparatus and the various numbering systems, could do no more than outline the position of Sinaiticus in relation to the wider tradition. What comes out clearly is the incomplete and uncontrolled character of these features: they are not fully added, and, with regard to the Eusebian system, the end result is by no means reliable. The nature of the corruption of the Eusebian system changes from Gospel to Gospel and seems unlikely to have been taken from the same manuscript that provided the main text. The differences between the Gospels may be an argument that we are dealing with corruptions that occurred during the transmission history of the system; it cannot be explained by differences on the scribal level.

4 SINGULAR READINGS AND THE INDIVIDUAL SCRIBAL PATTERNS OF SCRIBES A AND D

INTRODUCTION In the previous two chapters, we looked at both the interaction between the scribes and their respective individualities. We have seen that it is justified to study Sinaiticus at the level of the scribe as each scribe shows characteristic patterns in their habits. In this chapter we will go one step further and ask the question of whether the individuality of the scribe is detectable at the level of the text itself. Is it possible to distinguish individual scribal patterns in the copied text? The tool we will use in this chapter is the study of singular readings. A singular reading is practically defined as a reading for which no outside Greek support can be found. By applying this method to specific passages in Sinaiticus, we will be able to compare the work of scribe A with that of scribe D.

METHODOLOGICAL CONSIDERATIONS Past scholarship and the study of singular readings Over the last two centuries, editors who published a biblical manuscript have often given attention to the unique readings of their manuscript. Fewer scholars have actively sought to use singular readings as a tool for analysing the work of the scribes. In the following pages we will limit ourselves to a review of these contributions.

The nineteenth century: Tischendorf and Westcott-Hort In the 1863 preface to the New Testament edition of Sinaiticus, Tischendorf explains why even the readings that occur only in his manuscript still need to be taken very seriously: Only examples of readings of the third category remain, which Sinaiticus upholds without any of its allies. In these there is much distinction of 131

132

Singular Readings and Individual Scribal Patterns goodness. They are after all the readings which against the will of all are recommended with the highest probability; again, others have no commendation unless they are written in an old codex. Also in this matter a remarkable likeness of Sinaiticus with the Vatican codex exists. Because of the singular excellence of these books, of which it is hardly dubious that they retain the truth being so often approved by one and a second ally—and not seldom the approval of the old fathers shows, being abandoned by all other ancient codices—, it surely appears fitting that we should not dare to reject something, should they prove to be alone.1

Tischendorf argues that singular readings that occur in an old manuscript that has proven itself in so many cases where outside support was available should be highly valued. It is not clear in this context whether Tischendorf speaks about each singular reading, or whether at this stage he has already excluded those readings that are assignable to clear errors in the copying process. Scrivener, in his collation of Sinaiticus, also discusses readings where Sinaiticus stands alone.2 After noticing the value of those readings for which patristic evidence exists, Scrivener remarks that “for many readings Sinaiticus has no support; some of these may possibly be arbitrary alterations.” He then follows this with a brief list of possible examples.3 Westcott and Hort have a more qualified approach to singular readings in the discussion of their two favoured codices, Vaticanus and Sinaiticus.4 Writing entirely from the framework of the quest for the original text, they point out that the authority of a singular reading in one manuscript may stand against a shared reading by a whole host of other manuscripts, provided that this last group had a single, common original. Singular readings may be inherited from the exemplar or be simply the product of recent corruption. It is necessary to sift away the singular readings that are mere individualisms and cannot be singular survivals of the autograph text. Individualisms may have come into being by anything from clerical errors to alterations of mental origin. The clerical errors, in particular, enable the scholar to estimate the general level of accuracy attained by the scribe and provide a catalogue of the kinds of errors to which he was prone. C. Tischendorf, Novum Testamentum Sinaiticum (Leipzig: F.A. Brockhaus, 1863), xxxviii (own translation). 2 F.H.A. Scrivener, A Full Collation of the Codex Sinaiticus with the Received Text of the New Testament (Cambridge: 1864), l–lii. 3 ibid., li. 4 Westcott-Hort, 230–50. 1

Methodological Considerations

133

Knowledge of the habits of a scribe is useful in dealing with individual variations. Singular readings which make good sense and are not explained by clerical error or known scribal tendencies are to be considered more seriously. Often it is not easy to tell whether the individualisms are due to transcriptional error in the last stage or stages of transmission or whether these are ancestral. The text as it stands is always a product of both these factors. Westcott and Hort then go on to describe the individualities of Vaticanus.5 The scribe did not attain a high standard of accuracy, but his slips are not proportionally numerous or bad. Omissions because of returning to the wrong place in the exemplar do occur, often explainable by homoeoteleuton. More rarely one finds reduplications of this sort. The doubling of short words or letters, or the omission of repeated words or letters is more characteristic. Westcott and Hort disagree with the supposition that Vaticanus displays a preference for the omission of nonessential words. They think it rather more likely that Vaticanus shows a text free from such interpolations, as such words “are peculiarly liable to be inserted.”6 The few remaining individualisms are due to assimilation to neighbouring clauses, verses, or parallel passages. Occasionally one encounters a transposition. Westcott and Hort conclude regarding the individualities of Vaticanus: The final impression produced by a review of all the trustworthy signs is of a patient and rather dull or mechanical type of transcription, subject now and then to the ordinary lapses which come from flagging watchfulness, but happily guiltless of ingenuity or other untimely activity of brain, and indeed unaffected by mental influences except of the most limited and unconscious kind.7

Westcott and Hort then turn to what they call the “singular readings proper,” which are all those singular readings that deserve consideration to be accepted in the text. These singular readings of Vaticanus are described as being of high excellence and make a good claim to be genuine. This is confirmed by the subsingular readings of the manuscript.8

Westcott-Hort, 234. Westcott-Hort, 236. 7 Westcott-Hort, 237. 8 Subsingular readings are defined by having “only secondary support, namely, that of inferior Greek MSS, of Versions, or of Fathers, or of combinations of documentary authorities of these kinds.” Westcott-Hort, 230. 5 6

134

Singular Readings and Individual Scribal Patterns

Turning to Sinaiticus, Westcott and Hort are confronted with a different situation.9 Hardly any omission or repetitions of small groups of letters can be found, but the lapses due to rapid and careless transcription are more numerous, as are the number of substitutions. Many of these are individualisms of the scribe himself, some are doubtless older, either of Western or Alexandrian origin. The character of the singular readings gives little encouragement to look favourably upon these readings. A few of the subsingular readings of Sinaiticus may possibly or even probably be genuine. Westcott and Hort explain the relationship between Sinaiticus and Vaticanus as one in which Sinaiticus, like Vaticanus, goes back to a text near to the autographs but has come under a different set of influences than Vaticanus. As the above discussion shows, Westcott and Hort are very much aware of the phenomenon of individualisms and singular readings. They draw a sharp distinction between these two categories as their interest lies solely in the recovery of the original text; singular readings are to be considered, individualisms are to be ignored. Westcott and Hort bring their presuppositions to the actual study of the individual manuscripts; they do not hesitate to talk about “inferior” manuscripts, “excellent” readings, and “genuineness.” The individual manuscript is only studied in order to estimate its value for the reconstruction of the original text.

The twentieth century: Colwell and Royse In 1965, Ernest Colwell published an essay entitled “Scribal Habits in Early Papyri: A Study in the Corruption of the Text” in which he aims “to gain knowledge of an individual scribe’s habits, and thus to increase skill in the evaluation of that manuscript.”10 Colwell also believes that this will lead to a better knowledge of the habits of scribes in general and will increase skill in the evaluation of readings. This in itself leads to a better ability to establish Westcott-Hort, 246. E.C. Colwell, “Scribal Habits in Early Papyri: a Study in the Corruption of the Text”, in The Bible in Modern Scholarship (Nashville: Abingdon Press, 1965), 370– 89. Reprinted as ‘Chapter Eight: Method in evaluating scribal habits: a study of P45, P66, P75’ in idem, Studies in Methodology in Textual Criticism of the New Testament (Grand Rapids: Eerdmans, 1969). References are to the latter publication. Throughout his career, Colwell has always given much attention to method in the textual criticism of the NT. Barbara Aland praises the method of singular readings as developed by Colwell saying that it was developed ‘mit philologischer Einfühlung und disziplinierter Phantasie’. B. Aland, “Kriterien zur Beurteilung kleinerer Papyrusfragmente des Neuen Testaments”, in New Testament Textual Criticism and Exegesis, ed. A. Denaux (Leuven: University Press; Peeters, 2002), 1. 9

10

Methodological Considerations

135

kinship among manuscripts, free up the critical apparatus, and supports the use of external evidence in the scholar’s reconstruction of the original text. Colwell concentrates on the singular readings of three papyri, P45, P66, P72, and, pragmatically, defines a singular reading as any reading that has no support from Tischendorf’s eighth edition. He works with the assumption that all singular readings are the product of a scribe, realising that he will err on two fronts, namely by including readings that are not scribe-created but inherited from the exemplar and by excluding readings that find support in other manuscripts but are due to the same corruption occurring more than once. Colwell goes through three stages in studying these readings. First, he excludes itacisms from his list, but not without noting the general tendencies of each manuscript. Secondly, he separates nonsense readings from sensible readings. “The nonsense readings include words unknown to grammar or lexicon, words that cannot be construed syntactically, or words that do not make sense in the context.”11 Again, these readings are not simply ignored but used to learn something about the manuscript. The third stage is the classification of singular readings based upon their origin. Colwell distinguishes leaps from the same to the same, the omission of short words, harmonising (with special emphasis on harmonising to the immediate context), influence of similar forms, transpositions, and editorial changes. These categories are not applied rigidly but used to develop a description of each of the scribes of the papyri. Colwell formulates detailed conclusions on the scribes of the three papyri. On the basis of transpositions and omission of syllables and letters, Colwell concludes that “P75 copies letters one by one; P66 copies syllables, usually two letters in length; P45 copies phrases and clauses.”12 A little later, in the section on editorial changes, P45 is characterised as attempting “stylistic improvements” and aiming at “concise expression, at clarity, and at a good style.”13 Some of these characteristics are shared by P75, except “the passion for conciseness.” P66 shows a wildness in copying and some of his singular readings are due to clarification, to smoothing out the text, to simplification or to logical agreement with context.14 The first set of conclusions on the method of copying is rightly criticised by Klaus Junack.15 Using Dain’s Colwell, Studies, 111. ibid., 116. 13 ibid., 120. 14 ibid., 121. 15 K. Junack, “Abschreibpraktiken und Schreibergewohnheiten in ihrer Auswirkung auf die Textüberlieferung”, in New Testament Textual Criticism: Its 11 12

136

Singular Readings and Individual Scribal Patterns

model of copying, to which Colwell pays tribute as well, Junack assumes that a scribe will not copy a text that makes no sense to the scribe. Copying per single letter or per syllable breaks the text up into meaningless units. The occasional omission or doubling of syllables are more likely to be due to what Dain calls the fourth stage of the copying process, “jeu de main,” than to the actual copying procedure of the scribe. In a later study,16 Colwell looks back upon his method and, faced with the problem of distinguishing the scribe from his source, he qualifies the type of reading from which scribal habits can possibly be studied: the scribe’s contribution can be identified especially in nonsense readings or “in readings that have won universal rejection not only from other scribes but from all editors as well.”17 In practice, Colwell has returned to Westcott and Hort’s division of unique manuscript readings into “individualisms” and “singular readings,” only with a different terminology. However, in comparison with Westcott and Hort, Colwell succeeds in adding much detail to the study of the individualities of the scribe and therefore the resultant profile of the habits of a scribe is much more fine-tuned than, for example, the description of the scribe of Vaticanus by Westcott and Hort. One may wonder, though, to what extent it is justified to ascribe specific motives to a copyist on no other basis than that of variant readings. When, for example, the scribe of P75 is described as having a “passion for conciseness” one runs the risk of descending into overinterpretation and a speculative type of psychology. Colwell’s study was followed by the dissertation of J.R. Royse which is presented as a continuation of Colwell’s approach. Royse studies the singular readings of P45, P46, P47, P66, P72 and P75, and offers an introductory chapter in which the underlying theory of the study of singular readings is discussed.18 After noting that in past research, extensive appeal has been made to scribal habits in order to explain certain variant readings, Royse complains that, though it is admittedly true that scribes were able to make virtually any kind of error imaginable, studies of the likelihood of the Significance for Exegesis, ed. E.J. Epp and G.D. Fee (Oxford: Clarendon Press, 1981), 288–89. See also Chapter 1, page 23. 16 E.C. Colwell, “Hort Redivivus: A Plea and a Program”, in Transitions In Biblical Scholarship (Chicago: University of Chicago Press, 1968), 131–55. Reprinted under the same title in Colwell, Studies, 148–71. Again, references are to the latter. 17 Colwell, Studies, 161 and n. 3. 18 J.R. Royse, “Scribal Habits in Early Greek New Testament Papyri” (doctoral thesis, Graduate Theological Union, 1981), Chapter I ‘The Study of Scribal Habits,’ 8–87.

Methodological Considerations

137

occurrence of certain readings are sparse. Colwell’s method of studying singular readings provides a window into the activity of the scribe and bypasses the pitfall of relying on appeals as to how a scribe must have behaved. Also a justification is given for concentrating on the early, pre300, papyri only. Again basing himself on Colwell, Royse envisages a scenario in which the manuscript tradition of the New Testament moved from an uncontrolled to a rigorously controlled tradition. Scribal habits established for the fourth century may not provide reliable indications for the pre-300 period of transmission. Each reading of a manuscript can be seen not only as authentic or inauthentic, but also as correct or incorrect, that is, a faithful copy of the exemplar or a wrong copy. The correct readings take us behind the given manuscript, the incorrect readings provide a window on the scribe. Not all incorrect readings will show up among the singular readings, which are readings without any attestation from other manuscripts. Some of the incorrect readings may coincide with scribal creations of other scribes or may be a correction of an evident error in his Vorlage. Other scribal creations may have been copied into other documents and thus lose their singular status. Royse admits that “there seems to be no way in advance of the sort of study undertaken here to determine just when such coincidence or textual connection should be assumed, and so we have no method to distinguish the non-singular scribal creations from the readings copied correctly from the scribe’s Vorlage.”19 On the other hand, not all singular readings are created by the scribe. Royse cites Nestle and Hort who both argued that the authentic reading may have been preserved in only one manuscript and that such an authentic reading is therefore a singular reading which is not scribe-created.20 Royse deems this a very improbable scenario and argues that even if true in some cases, it would not affect the conclusions as to general patterns. However, one might object further that a reading does not have to be authentic in order to be part of a textual tradition that only left traces in a single manuscript. Royse counters this objection with three observations. First, because the NT has been transmitted by a tradition which is highly contaminated and which has left such vast quantities of manuscript evidence, it is likely that there have been only very few, if any, real dead ends within the manuscript tradition. Secondly, Royse points out that many scribal errors appear to have been readily corrected by later copyists. Though this observation was made by Colwell and Tune concerning nonsense readings, Royse argues that likewise 19 20

ibid., 32. ibid., 33.

138

Singular Readings and Individual Scribal Patterns

many sensible readings must have appeared obviously wrong to most scribes. Thirdly, for most purposes, it does not matter much whether the singular readings are the result of a manuscript tradition or the work of an individual scribe. The formulation of the scribal tendencies of a single manuscript would remain the same, only our assessment of the accuracy of the responsible scribe would have been skewed in the sense that we attribute habits to the scribe which are actually the accumulated errors of a whole tradition. All in all, Royse is much more positive than Colwell when it comes to the scribe-created nature of singular readings and disagrees with Colwell’s partial retraction from his earlier position. Royse establishes the singular status of a reading on the basis of the evidence presented in Tischendorf, von Soden, Nestle-Aland, and the United Bible Societies; versional evidence is mostly excluded except in cases where unambiguous support exists. Like Colwell, the definition of what establishes a singular reading is a pragmatic one, but unlike Colwell, Royse presents all the data on which he bases his analysis. Colwell considers the singular readings before the corrections by the original scribe, Royse chooses the readings after such corrections, though in an appendix he deals with these corrections in order to confirm the already established scribal habits. Also, Royse completely ignores some common orthographic variations in his collation, such as interchanges of ei–i, ai–e, oi–u, variations with regard to moveable nu, and breathings, accents, punctuation, and iota adscript. Royse deals with the remaining orthographic singulars only briefly, and also distinguishes between the strictly nonsense readings and the nonsense readings in context. The remaining group of singular readings are labelled the “significant singulars” and are subdivided using the formal categories of omissions, additions, transpositions, variations in grammatical forms, and differences in proper names. These categories suppose knowledge of the reading of the Vorlage. Royse proposes an intuitive method to reconstruct this reading taking account of the readings of the total tradition and of the known genealogical relationships of the document under consideration. In the actual analysis of the singular readings of the six papyri, Royse goes well beyond the purely formal categories and tries to understand the readings in relation to their causes. Royse presents as one of the chief outcomes of his study the fact that scribes omit more words than they add.21 He then contrasts this with the general attitude of textual scholars from Griesbach on, which is to prefer the shorter reading.22 Royse formulates even a revised canon based on 21 22

Royse, Scribal habits, chapter VIII ‘The Shorter Reading?’, 593–608. ibid., 593. Royse cites only part of Griesbach’s first canon.

Methodological Considerations

139

transcriptional probability that the longer reading is to be preferred unless 1) the longer reading appears to be late on genealogical grounds, 2) the longer reading can be explained on the basis of harmonisation to the immediate context, to parallels, or to general usage, and 3) the longer reading may have arisen from an attempt at grammatical improvement. Royse reiterates his earlier point that there is likely to have been a major shift in scribal practise in the fourth century. The canonisation of the biblical books as Scripture must have had an impact in the way these books were copied. Their increased status will have had a positive effect on the faithfulness of the copying process. Though Royse may be correct in seeing the history of reception of Griesbach’s canon as one that tends to repeat a simplistic interpretation of lectio brevior lectio potior, it is incorrect that Griesbach did not qualify his first canon. Perhaps Royse comes close to misrepresenting Griesbach by quoting only that part of the canon in which Griesbach mentions the rationale for preferring the shorter reading over the more verbose reading. The second part of the canon, summarised by Royse as mentioning “certain exceptional cases when the longer reading may be preferred (e.g. when the omission may have arisen by homoeoteleuton),”23 mentions six exceptions, thus qualifying what sort of readings Griesbach had in mind when he formulated the canon in the first place. The third exception to Griesbach’s canon is “if that which is lacking could be lacking without harming the sense or the structure of the sentence, as for example incidental, brief propositions, and other matter the absence of which would be scarcely noticed by the scribe when re-reading what he had written.”24 It is not surprising that the majority of omissions Royse noted in the six papyri fall under this category. Apparently Griesbach was only concerned with the few substantial rewritings and not with the vast majority of inconsequential readings. On the other hand, the reception of this canon in later scholarship shows that it is very easy to misunderstand Griesbach on this particular point.25

ibid., 594. Griesbach as cited in B.M. Metzger, The Text of the New Testament 3rd ed. (New York; Oxford: Oxford University Press, 1992), 120. 25 Griesbach: ‘... for scribes were much more prone to add than to omit. They scarcely ever deliberately omitted anything, but they added many things; certainly they omitted some things by accident, but likewise not a few things have been added to the text by scribes through errors of the eye, ear, memory, imagination, and judgement’ (in Metzger, Text, 120). Griesbach gives the false impression that 23 24

140

Singular Readings and Individual Scribal Patterns

In two articles, Peter Head applies Colwell’s and Royse’s method to the fragmentary papyri of the New Testament.26 In both these studies the general conclusions of Royse are vindicated: omissions are more common than additions, and the influence of harmonisation is relatively large. Also Barbara Aland has worked on the early fragmentary papyri and their singular readings.27 However, in contrast to Peter Head, she emphasises the importance of paying attention to all the variants found in the papyri after collating these with the hypothetically reconstructed starting point of the whole textual tradition, which is the NA27 text. One cannot exclude the possibility that variants arose independently in other manuscripts and therefore, in the case of the early fragmentary papyri, all the variant readings must be taken into consideration.28 Philip Comfort uses singular readings to describe a scribe’s role as a reader of the text.29 The singular readings in the early papyri are used as indications of how the scribe responds to the text, described in terms of reader reception analysis. Whether this approach really leads to new results is not clear yet. The method of studying singular readings and Codex Sinaiticus From the previous section it is clear that Colwell is the true father of the systematic study of singular readings. He approaches the whole phenomenon almost organically, refusing to analyse the readings in sharply defined formal categories, as is the basis of Royse’s analysis. Also, Colwell does not dismiss frequent orthographic errors, as Royse does, but uses accidental additions outnumbered accidental omissions though his exact wording as to the relative proportion of these two phenomena is ambiguous. 26 P.M. Head, “Observations on Early Papyri of the Synoptic Gospels, Especially on the ‘Scribal Habits’”, Biblica 71 (1990): 240–247, and idem, “The Habits of New Testament Copyists: Singular Readings in the Early Fragmentary Papyri of John”, Biblica 85 (2004): 399–408. 27 B. Aland, “Das Zeugnis der frühen Papyri für den Text der Evangelien: diskutiert am Matthäusevangelium”, in The Four Gospels 1992, ed. F. Van Segbroeck, et al. (Leuven: University Press; Peeters, 1992), 325-35. See also idem, “Kriterien”, 1–13. 28 A good example is the comment on the transposition of autwn koinwnoi 77 in P (Matt 23:30): ‘Ob der papyrus gar die Umstellung erfunden hat, lässt sich nicht sagen. Wahrscheinlicher ist, dass er sie von (einer) Vorlage(n), auf die auch seine Mitzeugen letzlich zurückgehen, zuverlässig kopiert hat. Auch unabhängiges, mehrfaches Entstehen ist denkbar.’ B. Aland, “Kriterien”, 5. 29 P. Comfort, “Scribes as Readers: Looking at New Testament Textual Variants According to Reader Reception Analysis”, Neotestamentica 38 (2004): 28– 53.

Methodological Considerations

141

them to gain a general idea of the scribe’s spelling ability. Another point of difference is the treatment of corrections. Royse starts with the corrected text, while Colwell uses the uncorrected text. These differences in approach may reflect different concerns on the part of these two scholars. While Colwell is genuinely interested in the scribal habits as part of the actual copying process, Royse is content with describing the patterns of the manuscript, regardless of whether these patterns are all due to the scribe who wrote a particular manuscript or whether the pattern is the result of an accumulation of errors. In the studies that have been carried out on the papyri so far, it has been assumed that the collective group of singular readings approximates the work of the individual scribe; some readings are included falsely because they are already present in the exemplar, other readings are excluded falsely as other scribes had made the same mistake independently in one or more other manuscripts. Colwell believed that these two errors would balance one another out. To Royse it does not make much difference; he concentrates rather on the features of a particular manuscript rather than on individual scribal activity. For our study of Sinaiticus, we will retain the elements from Colwell and Royse that are most suited to our interest. We will present all the evidence, and wherever possible we will work with the uncorrected text without ignoring the corrections. Besides this, it will be shown that Sinaiticus can also be fruitfully used to verify the method as such. In the following studies we will apply the study of singular readings to selected passages, and, for the first time, we will be able to actually check the method and test its validity. We will do this by looking at four sections. In 1 Chronicles we will look at a number of readings that are creations of the scribe and which were subsequently corrected by him or his colleague. These scribe-created copying errors are not necessarily singular readings but may occur in other manuscripts as well. They will give us an insight into the number and type of readings which arise independently in other manuscripts. The second section deals with the book of Psalms, which is written by two scribes, each of them responsible for a substantial part. Differences in the number and pattern of singular readings in the two parts are unlikely to reflect the exemplar but rather show us actual differences between scribe D and scribe A. Likewise in the NT, scribe D is responsible for the text of six folios within the work of scribe A. We will look at two of these replacement folios in the Pauline corpus and compare the singular readings with the surrounding text by scribe A, and one folio in Luke, which will be set against the following folios of this Gospel. By comparing these two scribes

142

Singular Readings and Individual Scribal Patterns

in various settings we will be able to explain variations in the number and character of singular readings in terms of their scribal habits rather than by reference to their exemplars. The main emphasis of this chapter will be on the work of scribe A, who is responsible for parts of the LXX and almost the whole NT. The second scribe of the NT, scribe D, will be discussed mainly in contrast with scribe A. Scribe B, the third scribe, will be ignored. The main reason for this is that it is not possible to compare his contribution with that of another scribe who worked from the same source material, as is the case with the other two scribes. We have noted before that his work seems to have been executed in relative isolation from the other two scribes. System of classification of singular readings Classification of singular readings should be no more than a tool in studying these readings and cannot replace a qualified discussion of the individual readings. However, in order to come to grips with the large amount of data, some sort of classification is called for. The following system is based on Colwell and Royse. The first category consists of orthographic singular readings. Orthographic changes of the vowel pairs ai–e and ei–i have already been discussed in the previous chapter, which, together with the other studies in that chapter, have provided us with an initial profile for each scribe. Then a discussion of nonsense words follows. The next two categories are leaps from the same to the same and the addition or omission of short words. Both categories of scribal errors belong to a basic level of copying. A problem in the categories of leaps is how to deal with leaps that cause a large portion of text to fall away or to be duplicated without the presence of an obvious “from the same to the same” element. In this study these have been classified under the addition and omission of clauses, except for the omission of whole poetic lines in the book of Psalms. Though there is not always a jump from the same text to the same text, we will regard the omission of a whole poetic line as a jump from line-ending to line-ending, which justifies its inclusion at this stage. This may have happened in prose texts as well, but there we do not have access to the physical line length of the exemplar. With the category of addition and omission of short words, harmonisation could, of course, play a role, but because of the very large uncertainty that would affect any conclusion on such phenomenon it seems simpler to assign all such readings to a single,

Methodological Considerations

143

formal category.30 Our definition of short words includes conjunctions, pronouns, articles, particles like an and eti, and the preposition en before dative constructions.31 To avoid confusion between the normal meaning of “short words” and our technical and pragmatically defined use, we will use the term verba minora to describe this category. As Colwell emphasised, harmonisations are an often-recurring scribal phenomenon. We have tried to distinguish between harmonisations to the immediate context, to the intermediate context, and to general usage. The immediate context is defined as the four or five verses immediately before or after the reading, the intermediate context as anything up to one or two chapters before or after the reading, and harmonisation to general usage as adaptation to common words or phrases, or to well known passages. Examples of each category can be found in the discussion below. The category of “editorial” rewritings comprises four subgroups: the meaning of one word can trigger a change in another word; the syntax of a sentence can be misunderstood so that various phrases are brought into a connection with one another which, on a correct reading, should not have taken place; errors may arise because of a misunderstanding of a correction in the exemplar; and difficult or unusual textual constructions may be smoothed out. After these readings, those forms are discussed that do not make sense in the context in which they stand. Many of the readings in this category may be no more than scribal slips and are not different in character from nonsense words. The advantage of discussing these words at this stage is that harmonisations and editorially inspired readings have already been filtered out. The remaining readings are then described formally as substitutions, transpositions, addition and omission of words and clauses, and major rewritings of the text. The various categories have some overlap with one another and are a mixture of formal and interpretative categories. This is in line with the method of Colwell, who tried to concentrate on the cause of scribal changes rather than on a purely formal description. Admittedly, grouping all the readings in purely formal categories would ensure greater perspicuity, but it would also be less instructive.

30 Compare A. Dain, Les Manuscrits 3rd ed. (Paris: Societe d’Edition Les Belles Lettres, 1975), 48. 31 The decision to include only this preposition is made on the basis of the frequency with which it is added and omitted.

144

Singular Readings and Individual Scribal Patterns

SINGULAR READINGS AND CORRECTIONS IN 1 CHRONICLES Introduction Sinaiticus preserves approximately 10 chapters of 1 Chronicles. These chapters precede, without any clear break, the text of 2 Esd 9:9 and it seems clear that the text of 1 Chronicles is a large intrusion into 2 Esdras. Though the beginning of this intruded section is lost, we may assume that it was regarded by the scribes as a genuine part of the text of 2 Esdras. This misplaced section of 1 Chronicles was, like the rest of 2 Esdras, corrected by the earliest scribes, and as the correctors did not note the intrusion of 1 Chronicles in 2 Esdras, it follows that they compared Sinaiticus against a codex with exactly the same intrusion. As it is unlikely that such an error would go unnoticed for long, we may assume therefore that the manuscript with which Sinaiticus was compared for the corrections was the same manuscript from which Sinaiticus was copied in the first place.32 If this is true, it means that each correction is a correction of an error the scribe created whilst copying his text. Though it would be ideal if this group of corrections could be used to identify all scribe-created errors in the text, it can only be used to see which readings were corrected. One cannot assume that each type of scribecreated error was corrected with the same frequency; some may have escaped the attention of the corrector altogether, and others may have been such apparent errors that even without the help of the exemplar the error could be corrected. Anyway, all the scribal errors identified by means of the fact that they were corrected give us an insight into the types of scribecreated readings. We will compare these corrected readings with the total number of singular readings that remain in the text and which may or may not have been copied from the exemplar. As indicated earlier, the study of the corrections in the preserved section of 1 Chronicles serves in our verification of the method of using singular readings. Both Colwell and Royse were aware of the two imprecisions in using singular readings as a source for scribe-created readings: readings that originated independently in other manuscripts are falsely excluded, and singular readings that are inherited from the parent are falsely included. In normal circumstances it is very hard to get an idea about the scale of these two errors. Our group of corrected scribal errors can possibly help to gain an insight into the first of these two errors. Those readings that were corrected by the scribes but do have support from other 32

Milne-Skeat, 2.

Singular Readings and Corrections in 1 Chronicles

145

manuscripts must have come into existence by independent reduplication of that same error in other manuscripts. Contamination of readings because the scribe knows the text almost by heart is not very likely in the case of 1 Chronicles, with the possible exception of 1 Chr 16 and 17.33 The relatively unfamiliar nature of 1 Chronicles, confirmed by the failure to note the intrusion, is an advantage for our study. However, the large number of nonGreek proper nouns may distort the overall picture and we will deal with these separately. The text and the correctors The exact content of the intruded section of Sinaiticus as it is preserved nowadays is 1 Chr 9:27–19:17 and amounts to 4.94 folios. On two of the five folios that contain this text, the running title esdras b occurs.34 As the folios before 1 Chr 9:27 are lost, we cannot tell anything about the start of this section, but the transition to the actual Esdras text is informative. The text at the critical point reads as follows: 1 Chr 19:17, folio 35.4 (OT 7), line 8.25–27 -as da+d kai epole mhsen auton ks+ o qs+ hmwn kai e-

In line 26 the word auton finishes 1 Chr 19:17. The next word on the line, ks+ is the start of the 2 Esdras text, and resumes somewhere within 2 Esd 9:9. In the lower margin, under the last column of the page, a note is written by a later hand, which is closely related to that of the C-correctors: mecri tou shmeiou twn triwn staurwn estin to telos twn epta fullwn twn perisswn  mh ontwn tou es dra 33 1 Chr 16 contains parts of Pss 105, 96 and 106; 1 Chr 17 (parallel to 2 Samuel 7) contains the promise of God to David. Both texts could have been popular in the early church because of their Messianic application. See e.g. Justin, Dial. 118. 34 Folio 35.1 (OT 4) recto and folio 35.3 (OT 6) recto.

146

Singular Readings and Individual Scribal Patterns

Milne and Skeat give the following translation: “At the sign of the three crosses, is the end of the seven leaves which are superfluous and not part of Esdras”.35 We find these three crosses placed at the bottom of the column (line 48) instead of next to line 26. However, in the margin before line 26, one or two partially erased crosses are still visible. Milne and Skeat observe rightly that the current note is not by the hand of corrector Ca and must be later, though it still belongs to the seventh or eighth century. What remains visible of the first version of the note is enough to show that it does not come from the hand of any of the earliest correctors and must come from after the scriptorium stage. As the original place of the one or two crosses has been erased, this is the second version of the note. Milne and Skeat suggest that the reason why the note and the crosses were replaced might have been that the note spoke of seven leaves and was, apparently, later interpreted as seven full leaves. The intruded text of 1 Chronicles is not touched by any of the Ccorrectors, unlike the subsequent text of 2 Esdras. If the earliest correctors had used a second manuscript of 2 Esdras they would have noticed the misplacement easily.36 The later correctors left the inserted text of 1 Chronicles untouched and simply noted the error because their copies lacked this part of 1 Chronicles in the middle of 2 Esdras. In total there are 90 corrections to the text, some of which can be ascribed with certainty to scribe A or scribe D, while the remaining corrections may have been made by either scribe. These corrections are more or less evenly distributed over the five folios and amount to an average of 2.1 corrections per column.37 The identity of the corrector who is responsible for the corrections is hard to establish in most individual cases. The large correction above folio 34.8 (OT 3), column 8, is undoubtedly by scribe D,38 as is the addition of kai ozeihl four folios later.39 Some of the smaller corrections are unlikely to be from the same hand and are more likely to be by scribe A, but it is hard to establish Milne-Skeat, 2. Tischendorf, Codex Friderico-Augustanus, 10. The only alternative possibility is the unlikely event of a correction against a manuscript which also derived from the same imperfect exemplar. 37 For comparison: the Gospel of Luke has approximately 192 corrections by the prima manus and early correctors in 155 columns (1.2 per column), Acts has 61 corrections in 152 columns (0.4), Revelation 74 corrections in 69 columns (1.1). I have no explanation for the relatively high number in 1 Chr. 38 Milne-Skeat, figure 12. 39 1 Chr 16:6, folio 35.3 (OT 6), line 2.22. 35 36

Singular Readings and Corrections in 1 Chronicles

147

whether these were made whilst copying or in a separate correction session. In addition to the analysis of the corrections, we have also made a listing of all the uncorrected singular readings of the 1 Chronicles section. We found 147 singular readings, which are defined as having only attestation from Sinaiticus in Brooke-McLean, which is, to date, still the only critical edition that provides an apparatus. These are all listed in Appendix IV: “Uncorrected singular readings in 1 Chronicles” on page 283. For both the uncorrected singular readings and the corrections we have tried to establish the likely text of the exemplar. In practice this was done by choosing that manuscript reading that diverged the least from the reading in Sinaiticus. First Chronicles is a relatively poorly attested part of the Greek Bible. Brooke-McLean quotes only 25 manuscripts, some of which contain only a part of the book.40 This means that our knowledge of the various textual clusters is limited and that it may be that a considerable number of singular readings are actually part of a subsection of the textual tradition rather than individual creations by the scribe who wrote the text. Though we will work with the singular readings of 1 Chronicles, we can be less sure about the true singular character of these readings than in the better-attested parts of the Greek Bible, which will be discussed later in this chapter. Discussion of the evidence

Orthography ℵ 1 Chr 9:37, folio kamakel34.8 (OT 3), lwq line 2.14 anacwnei 1 Chr 11:29, folio 35.1 (OT 4), line 1.44 faraqwqei 1 Chr 11:31, folio 35.1 (OT 4), line 2.2

correction kai makellwq

exemplar -

acwnei

acwi

faraqwnei

faraqwn(e)i

40 Brooke-McLean, v. Detlef Fraenkel, in the new Verzeichnis, gives only five manuscripts of 1 Chr from the eight century or earlier: Sinaiticus (S), Alexandrinus (A), Vaticanus (B), P. Sinai Gr. 1 (880), and Codex Venetus (V) (D. Fraenkel, Verzeichnis der griechischen Handschriften des Alten Testaments [Göttingen: Vandenhoeck & Ruprecht, 2004], 478).

148

Singular Readings and Individual Scribal Patterns

1 Chr 11:33, folio 35.1 (OT 4), line 2.6 1 Chr 11:43, folio 35.1 (OT 4), line 2.32 1 Chr 11:44, folio 35.1 (OT 4), line 2.35 1 Chr 12:4, folio 35.1 (OT 4), line 3.22 1 Chr 12:5, folio 35.1 (OT 4), line 3.29 1 Chr 12:21, folio 35.1 (OT 4), line 5.33 1 Chr 15:18, folio 35.2 (OT 5), line 7.29 1 Chr 15:24, folio 35.2 (OT 5), line 8.22 1 Chr 15:29, folio 35.3 (OT 6), line 1.17 1 Chr 16:1, folio 35.3 (OT 6), line 1.26 1 Chr 16:5, folio 35.3 (OT 6), line 2.15 1 Chr 17:1, folio 35.3 (OT 6), line 6.29 1 Chr 18:5, folio 35.4 (OT 7), line 3.13 1 Chr 18:12, folio 35.4 (OT 7), line 4.29

ℵ beerbein

correction beermein

exemplar -

iwsafas

iwsafat

-

eia

ieia

-

agabwniths gabawniths

-

ieremihas

ieremias

-

swkag

swka

swklag ?

eliba

eliab

-

eia

ieia

i(e)ia

melcorl

melcol

-

eisenegkan

eishnegkan

-

benias

banias

banaias ?

aqan

naqan

-

adraza

adrazar

-

allwn

alwn

-

Singular Readings and Corrections in 1 Chronicles

1 Chr 18:16, folio 35.4 (OT 7), line 5.3 1 Chr 19:10, folio 35.4 (OT 7), line 7.13 1 Chr 19:16, folio 35.4 (OT 7), line 8.11 1 Chr 19:16, folio 35.4 (OT 7), line 8.13

ℵ abieqer

correction abiaqar

exemplar -

gegonan

gegonasin

-

eswfar

eswfac

swfac ?

edraazar

adrazar

-

149

Twice an orthographic correction took place on verbs: eisenegkan is corrected back to the properly augmented eishnegkan, and the form gegonan is corrected to gegonasin.41 Two uncorrected orthographic readings both involve the augmented syllable oi. Scribe A maintains oi while the Rahlfs text has w//.42 Admittedly, the difference is only slight, but as the orthography is confirmed in Brooke-McLean it warrants inclusion. Slightly more rare in biblical manuscripts is the uncorrected pin for piein. This form is a common phenomenon according to Gignac.43 All the remaining corrections and singular readings in this category concern proper nouns. Nineteen times a proper noun is corrected, three of which are still singular readings after correction. The uncorrected singular readings of names number 56.44 In several places in our section of 1 Chronicles, individual names in a long list are written with each name starting a new line or with interpunction separating the various items. This makes it possible to see in which cases a singular reading was created by a unique division of syllables over two consecutive names.45 Sinaiticus has two unique combinations: isoba 41

55.

Both forms were interchangeable in non-literary papyri, see Gignac II, 354–

See also Gignac II, 239. Gignac I, 295. 44 At 1 Chr 15:24, Rahlfs adopts sobneia in the main text, which is only supported by Sinaiticus. Given the high frequency of orthographic singulars this may be questionable. 45 However, the singular status of such readings is only true in relation to other manuscripts with some means of indicating word division by interpunction or accentuation. 42 43

150

Singular Readings and Individual Scribal Patterns

amnaqan for isabaam naqan in 1 Chr 14:4, and samai anelhm for samaian enhl in 1 Chr 15:11. Another word division problem in names is the connection of a preceding kai with the proper noun. In kaeqema, the i has dropped out of kai eqema and was not corrected, but in kamakellwq the error was corrected.

Nonsense words

1 Chr 9:33, folio 34.8 (OT 3), line 1.39 1 Chr 11:2, folio 34.8 (OT 3), line 5.43 1 Chr 11:23, folio 35.1 (OT 4), line 1.10 1 Chr 14:8, folio 35.2 (OT 5), line 3.31 1 Chr 15:8, folio 35.2 (OT 5), line 6.6 1 Chr 15:28, folio 35.3 (OT 6), line 1.3 1 Chr 18:11, folio 35.4 (OT 7), line 4.19

ℵ

correction

exemplar

diatetagmen efhmeria

diatetagmenoi efhmeria

-

peneis

poimeneis

poimaneis ?

aoraton

oraton

-

hsousan

hkousan

-

c added

diakosioi

-

anagaontes

anagontes

-

eqnwnwn

eqnwn

-

Seven times we find a correction of words which are nonsensical in their original form. Three times, a nonsense word was created by adding a letter which rendered the word into a nonsense word. However, if the prima manus was responsible for the corrections, then it may well be that he corrected himself on the spot, so that the resultant text was correct. A good example may be the following variant: 1 Chr 11:23, folio 35.1 (OT 4), line 1.8–11 kai epataxen ton andra ton aigupti-

Singular Readings and Corrections in 1 Chronicles

151

on aoraton pentaphcun kai en

It is unlikely that the scribe really intended the word aoraton (“invisible”); rather we likely have a leap back to one of the words starting with a, which was noticed and corrected after the first letter was written down. Only two uncorrected nonsense words remain in the text. One involves a dropped syllable omati for onomati, not noted by the corrector.46 An extra omicron was inserted in omoroountes.

Leaps from the same to the same 1 Chr 12:22, folio 35.1 (OT 4), line 5.43

ℵ eisinemachsan

correction sunemachsan

exemplar -

One leap from the same to the same has been corrected. The scribe caught himself in the act at 1 Chr 12:22, where after writing eisin, taken from the previous verse, he corrected it into the first letters of sunemachsan.47 Two singular readings remain: the omission of kai zambri after kai ton zambri, and the omission of enantion sou as a result of a jump from sou to sou.

Addition/omission of verba minora 1 Chr 10:12, folio 34.8 (OT 3), line 5.12 1 Chr 11:4, folio 34.8 (OT 3), line 6.13 1 Chr 11:6, folio 34.8 (OT 3), line 6.25 1 Chr 14:1, folio 35.2 (OT 5), line 3.2

46 47

ℵ iabeis

correction en iabeis

exemplar -

andres ihl+m

andres autou ihl+m

andres autou eis ihl+m ? -

ton iebousaion iebousaion

omitted

kai

-

On the same line the corrector had already made a correction. Note that this is verse 21 in Brooke-McLean instead of 22.

152

Singular Readings and Individual Scribal Patterns

1 Chr 14:4, folio 35.2 (OT 5), line 3.20 1 Chr 15:26, folio 35.2 (OT 5), line 8.39 1 Chr 15:27, folio 35.2 (OT 5), line 8.40 1 Chr 16:30, folio 35.3 (OT 6), line 4.32 1 Chr 16:31, folio 35.3 (OT 6), line 4.35 1 Chr 16:43, folio 35.3 (OT 6), line 6.20 1 Chr 17:18, folio 35.4 (OT 7), line 1.15 1 Chr 17:19, folio 35.4 (OT 7), line 1.20 1 Chr 18:11, folio 35.4 (OT 7), line 4.15

ℵ hsan

correction oi hsan

exemplar -

epta

an epta

-

stolh

en stolh

-

kai katorqwqhtw

katorqwqhtw

-

o ouranos

ouranos

-

omitted

autou

-

omitted

eti

-

kai thn

thn

-

omitted

tauta

-

Thirteen times the addition and omission of verba minora is corrected, in nine cases this concerned the repairing of an omitted word. In 1 Chr 11:4 the problem was only partially mended, so that the resultant text still yields a singular reading. We have counted a further 16 singular readings where a short word is added or omitted: ten of these consist of omissions, the other six are additions.

Harmonisations Harmonisation to the immediate context ℵ 1 Chr 10:8, folio tas 34.8 (OT 3), line 4.23

correction tous

exemplar -

Singular Readings and Corrections in 1 Chronicles

1 Chr 11:9, folio 34.8 (OT 3), line 6.45 1 Chr 11:11, folio 34.8 (OT 3), line 7.16 1 Chr 11:20, folio 34.8 (OT 3), line 8.34 1 Chr 12:38, folio 35.1 (OT 4), line 7.48 1 Chr 12:39, folio 35.1 (OT 4), line 8.12 1 Chr 15:1, folio 35.2 (OT 5), line 5.17 1 Chr 16:4, folio 35.3 (OT 6), line 2.5 1 Chr 17:13, folio 35.3 (OT 6), line 8.29 1 Chr 17:21, folio 35.4 (OT 7), line 1.33 1 Chr 19:16, folio 35.4 (OT 7), line 8.7

153

ℵ poreuome-nos met autou

correction exemplar poreuome-nos -

triakosias

triakosious

-

triakosias

triakosious

-

fulous

fulhs

-

oi kataloipoi

o[inc]i[/inc] kataloipos

o kataloipos

thn kibwton

th kibwtw

-

litourgountwn

litourgountas

l(e)itourgountas

apesthsa

aposthsw

-

eauton

eautw

-

apestilen

apestilan

apest(e)i-lan

Eleven cases of harmonisation to the immediate context have been corrected; there remain 12 singular readings in the text that have not been corrected. The influence from the immediate context can be seen in a harmonised case-ending, for example tois eidolois autois (for autwn), and in repetition of a words from earlier in the verse, such as pas in pas is_l (2) in 1 Chr 10:7. Also it happens frequently that influence from further down in the text is felt, such as 1 Chr 11:9 where the phrase met autou is already inserted after poreuomenos.

154

Singular Readings and Individual Scribal Patterns

Harmonisation to general usage 1 Chr 17:9, folio 35.3 (OT 6), line 7.47 1 Chr 17:13, folio 35.3 (OT 6), line 8.28 1 Chr 17:13, folio 35.3 (OT 6), line 8.30 1 Chr 19:5, folio 35.4 (OT 7), line 6.4

ℵ mnhsqhsei

correction prosqhsei

exemplar -

laon

uion

-

autwn

autou

-

basilei da+d

da+d

-

Four cases of harmonisation to general usage are corrected. The addition of basilei to da+d was removed; the substitution of prosqhsei with mnhsqh sei was reverted in the phrase kai ou mnhsqhsei adikia tou tapinwse48 auton in 1 Chr 17:9; and likewise, a few verses further down at 17:13, laon was corrected to uion: egw esomai autw eis patera kai autos estai moi eis laon. Interestingly, both of the last two errors can be explained by influence from Jer 31 (LXX 38):33–34 or influence from the citation of these verses in Heb 8. One uncorrected singular reading was found in this category, which is an influence from general biblical usage: in 1 Chr 10:2 we find the well known melcisedek instead of melcisoue. Most cases of harmonisations are substitutions, but some of them are additions. Editorial rewritings 1 Chr 12:19, folio 35.1 (OT 4), line 5.12 1 Chr 10:7, folio 34.8 (OT 3), line 4.10 1 Chr 11:19, folio 34.8 (OT 3), line 8.21 1 Chr 11:19, folio 34.8 (OT 3), line 8.25

48

ℵ prosetaxato

correction prosedexato

exempl. -

on

o

-

touto

toutwn

-

hboulonto

ebouleto

-

Read tapeinwsai.

155

Singular Readings and Corrections in 1 Chronicles

1 Chr 14:14, folio 35.2 (OT 5), line 4.30 1 Chr 14:14, folio 35.2 (OT 5), line 4.31 1 Chr 9:38, folio 34.8 (OT 3), line 2.15 1 Chr 15:19, folio 35.2 (OT 5), line 7.40 1 Chr 17:21, folio 35.4 (OT 7), line 1.29

ℵ kai eipen added

correction eti

exempl. -

autw kai eipen

kai eipen autw

-

omitted

-

kai added

en mesw twn adelfwn autwn poihsai

eqnh

eqnos

-

-

Under the category editorial rewritings, nine initial mistakes can be found. In 12:19, the scribe wrote originally that David prosetaxato autous instead of prosedexato. The context is that of receiving the tributes of troops, which may have triggered the verb prostassw, helped by the fact that these words are close in pronunciation. An example of a misunderstanding of the syntax is found in 1 Chr 10:7, where the scribe initially wrote kai eiden pas is_l on en tw aulwni oti efugen ktl., apparently thinking that the o after is_l had to be a relative pronoun. It was correctly changed to the article o. Three examples of smoothing the text can be found in 1 Chr 9:38, 15:19 and 17:21. In the first example, the text of the exemplar probably read kai outoi en mesw twn adelfwn autwn katwkhsan en ierousalhm en mesw twn adelfwn autwn. Understandably, the second en mesw twn adelfwn autwn was removed because of the obvious repetition in the Greek text. However, as it was found in the exemplar, it was subsequently reinserted into the text. In the second example the scribe was confronted with two consecutive infinitives: tou akousai poihsai. This unusual construction was resolved by inserting kai, which solved the problem, but in the correction phase the kai was rightly removed. Among the uncorrected singular readings we find nine instances of editorial readings. Two of these are found in 1 Chr 14:15–16, where the text is presented as if David is the speaker rather than the two men who had got water from the well in Bethlehem, which is a misreading of the flow of the

156

Singular Readings and Individual Scribal Patterns

story. Six other examples may also show a misunderstanding of the syntax, though alternative explanations are certainly possible. One example of an editorial addition is found in 17:21. The phrase h ceir sou is added to the imperative pistwqhtw (“let your hand be established”). A proper subject was already mentioned earlier in the verse, o logos sou, albeit that it is quite far removed from the verb. The combination of the verb pistow with ceir does not occur elsewhere in the LXX.

Nonsense in context, substitutions, and transpositions Nonsense in context 1 Chr 10:13, folio 34.8 (OT 3), line 5.25 1 Chr 11:41, folio 35.1 (OT 4), line 2.27 1 Chr 12:2, folio 35.1 (OT 4), line 3.13 1 Chr 14:3, folio 35.2 (OT 5), line 3.15 1 Chr 19:14, folio 35.4 (OT 7), line 7.41

ℵ autw

correction exemplar aut[inc]o[i/ nc] autw ? autw

uiw

uios

-

kai

ek

-

eteqhsan

etecqhsan

-

efu

efugon

-

ℵ ihl+m

correction ih+l

exemplar israhl

arciereis

iereis

-

o swthr

ths swthrias -

egeneto

eneteilato

Substitutions 1 Chr 11:2, folio 34.8 (OT 3), line 5.46 1 Chr 15:14, folio 35.2 (OT 5), line 6.38 1 Chr 16:35, folio 35.3 (OT 6), line 5.7 1 Chr 16:40, folio 35.3 (OT 6), line 6.3

-

Singular Readings and Corrections in 1 Chronicles ℵ polin 1 Chr 16:42, folio 35.3 (OT 6), line 6.20 1 Chr 18:4, folio rhmata 35.4 (OT 7), line 3.8 1 Chr 18:4, folio eikosi 35.4 (OT 7), line 3.9

157

correction pulhn

exemplar -

armata

-

r_

ekaton

correction mou epi ton laon ‰

exemplar epi ton laon mou

Transpositions ℵ 1 Chr 17:7, folio mou epi ton 35.3 (OT 6), laon line 7.29

Five cases occur in which words that are grammatically correct in themselves but do not make sense in their context are corrected. Against these stand nine uncorrected cases. Also eight substitutions are corrected, leaving 26 singular readings in the text. There is only one corrected transposition, mou epi ton laon to epi ton laon mou. However, the corrector adds the second mou without removing the first one, thus treating this reading as an omission of mou instead of as a transposition. Five transpositions are left uncorrected.

Add and omit words and clauses 1 Chr 9:29, folio 34.8 (OT 3), line 1.14 1 Chr 9:32, folio 34.8 (OT 3), line 1.29 1 Chr 9:44, folio 34.8 (OT 3), line 3.1 1 Chr 10:10, folio 34.8 (OT 3), line 4.41

ℵ omitted

correction tou elaiou

exemplar -

omitted

kata sabbaton

-

kai ezreikan prwtotokos autou ismahl added kai thn autou kai thn kefalhn autou

-

-

158

Singular Readings and Individual Scribal Patterns

ℵ 1 Chr 11:2, folio omitted 34.8 (OT 3), line 5.45 1 Chr 11:18, omitted folio 34.8 (OT 3), line 8.12

1 Chr 12:16, folio 35.1 (OT 4), line 4.26 1 Chr 15:3, folio 35.2 (OT 5), line 5.32 1 Chr 15:25, folio 35.2 (OT 5), line 8.25 1 Chr 16:6, folio 35.3 (OT 6), line 2.22 1 Chr 16:10, folio 35.3 (OT 6), line 2.44

correction esh

exemplar -

omitted

dierrhxan oi g_ as correction thn parembolhn twn allofulwn  udreusanto udwr ek tou lakkou tou en beqleem os hn en th pulh kai ton iordanhn -

omitted

eis ton topon -

omitted

isl+

israhl

omitted

 ozeihl

kai oz(e)ihl

omitted

aineite

-

Ten times an omitted word or phrase is corrected, and once a substantial jump back is put right. The repetition of the words ezreikan prwtotokos autou ismahl, repeats the orthographic singular reading ezreikan. In an uncorrected jump back in 1 Chr 12:4, the orthographic singular agabwniths of the first version is avoided and the normal gabawniths is written down. However, in the first version agabwniths is corrected by prima manus or an early corrector. It may be that agabwniths was already put straight before the scribe wrote the word the second time. Besides this uncorrected addition, four uncorrected omissions remain.

Singular Readings and Corrections in 1 Chronicles

159

Correction strategy After this overview, what can be said about the type of mistakes that are corrected? Analytic Category Orthography - proper nouns - other Nonsense word forms Leaps a/o verba minora Harmonisation - immediate context - intermediate context - general usage Editorial readings - interfering meaning - misreading syntax - misunderstood corrections - deliberate improvements Nonsense meanings Substitutions Transpositions a/o words or clauses More complex rewritings Total:

Corrected readings/10 folios (real number) 42.5 (21) 38.5 (19) 4.0 (2) 14.2 (7) 2.0 (1) 26.3 (13) 30.4 (15) 22.3 (11) 8.1 (4) 18.2 (9) 2.0 (1) 10.1 (5) 6.1 (3) 10.1 (5) 16.2 (8) 2.0 (1) 22.3 (11) 182.2 (90)

Uncorrected singulars/10 folios (real number) 121.5 (60) 115.4 (57) 6.1 (3) 4.0 (2) 4.0 (2) 32.4 (16) 26.3 (13) 24.3 (12) 2.0 (1) 18.2 (9) 16.2 (8) 2.0 (1) 18.2 (9) 52.6 (26) 10.1 (5) 10.1 (5) 297.6 (147)

The itacisms ei–i and ai–e are not subject to corrections. Though in this section scribe D added corrections to the work of scribe A, and demonstrates elsewhere that he is a far better speller in the portions of Sinaiticus that he copies himself, none of these orthographic variants are touched. However, some form of morphological/orthographic awareness is evidenced in the correction of gegonan to gegonasin. The spelling of proper nouns is a weakness in the whole section, and the correction phase left many singular spellings intact. The fact that 19 corrections of proper nouns were made may suggest that many of the uncorrected singular spellings also arose during the copying process. In that case, the implication is that total orthographic purity was not aimed for in the copying process nor in the correction phase. As one would expect of a corrected text, not many nonsense words are left uncorrected. These words are probably the easiest to spot when one

160

Singular Readings and Individual Scribal Patterns

rereads a text. Remarkable is the high number of the omitted and added verba minora that are corrected: 12 corrected cases against 16 remaining singulars. The to-be-corrected cases show that the scribe is quite prone to this error; the corrections show that reasonably often these are detected during the correction phase. The same is true for the various harmonisations, which happen regularly, but are also often corrected. The scribe is not particularly prone to errors in word order. Only one case is corrected, five singular readings remain. The only corrected reading is treated as an omission rather than as a transposition. When it comes to the correction of omitted and added clauses or words, eleven have been corrected and only five singulars are left. Ten corrected omissions mean that the scribe regularly omits words, but it also shows that a high percentage is corrected. Relatively speaking, substitutions seem to have been much less frequently noticed and corrected, though the absolute number of corrections is higher than that of omitted and added clauses or words. Scribe created readings that are not singular readings. Having collated the uncorrected version of the readings that are corrected with the apparatus in Brooke-McLean, we can now determine how many of the scribe-created, and then corrected readings show up in other manuscripts as well. The following 13 readings are found in other manuscripts, indicating that elsewhere in the tradition a similar mistake was made: Orthography 1 Chr 15:24, folio 35.2 (OT 5), line 8.22 1 Chr 18:12, folio 35.4 (OT 7), line 4.29 1 Chr 19:10, folio 35.4 (OT 7), line 7.13

primary reading eia (c2)

corrected reading ieia (most others)

allwn (g)

alwn (all others)

gegonan (A c2)

gegonasin (all others)

primary reading iabeis (g m)

corrected reading en iabeis (all others)

ton iebousaion (N)

iebousaion (all others) an epta

verba minora 1 Chr 10:12, folio 34.8 (OT 3), line 5.12 1 Chr 11:6, folio 34.8 (OT 3), line 6.25 1 Chr 15:26, folio 35.2

epta (see discusion)

Singular Readings and Corrections in 1 Chronicles

(OT 5), line 8.39 1 Chr 15:27, folio 35.2 (OT 5), line 8.40

161

primary reading

corrected reading

stolh (A)

en stolh (all others)

primary reading apestilen (A N all others) autwn (c2)

corrected reading apestilan (B h c2)

touto (e e2)

toutwn (all others)

Harmonisation 1 Chr 19:16, folio 35.4 (OT 7), line 8.7 1 Chr 17:13, folio 35.3 (OT 6), line 8.30 1 Chr 11:19, folio 34.8 (OT 3), line 8.21

autou (all others)

Smoothing of the text 1 Chr 9:38, folio 34.8 (OT 3), line 2.15

primary reading omitted (c d e j m p q t z)

corrected reading en mesw twn adelfwn autwn (others)

primary reading omitted (e2)

corrected reading tou elaiou (all others) esh (all others)

Omissions 1 Chr 9:29, folio 34.8 (OT 3), line 1.14 1 Chr 11:2, folio 34.8 (OT 3), line 5.45

omitted (A N c e g i n)

The koine form gegonan for the classical gegonasin is found also in A and c2. The uncorrected spelling of the name ieia (eia) appears also in c2, yet with an orthographic change (ia), the spelling of alwn as allwn, an easy error to make, is found in g. In the category of adding/omitting verba minora, four of the thirteen corrected cases occurred in other manuscripts. One of these comprises a case in which Sinaiticus—if we take the whole variant unit—has a unique variant. The text written by prima manus is equsan epta moscous kai epta krious (all witnesses except B and c2 agree with prima manus), yet the second epta is corrected to an epta, which must have been the reading of the exemplar. The first epta was written by scribe A without an and, initially, he apparently harmonised the second epta to the first. This variant demonstrates the process by which the text with an epta twice,

162

Singular Readings and Individual Scribal Patterns

evidenced in B c2, develops to that of the other manuscripts, all reading just epta. Originally the first an may have been dropped out with the help of the preceding equsan, the second an followed because of harmonisation. Whether the exemplar of Sinaiticus read an at both places or just at the second epta cannot be established. In 1 Chr 19:16, an example of harmonisation to the immediate context is shared with the majority of the tradition and the corrected variant is found in three manuscripts, B h c2. In this case the primary reading of Sinaiticus, shared with the majority of the tradition, has the singular verb apesteilen, with the implied subject suros from the previous clause. This is, in Sinaiticus, corrected to apesteilan, where the plural may reflect the Hebrew.49 The obvious smoothing of the sentence by omitting the repeated en mesw twn adelfwn autwn is widely found (e d e j m p q t z). Two omissions of larger words or phrases are shared with other manuscripts: the omission of tou elaiou with e2 and the omission of esh with A N c e g i n. Of the 90 scribe-created and corrected errors, 13 arose independently in other manuscripts, which amounts to 14%. However, one should realise that this does not mean that 14% of all scribe-created errors are independently made in other manuscripts also, but only that 14% of the corrected errors also occur in other manuscripts. If, for the sake of the argument, we assume that all the uncorrected singular readings and the corrections taken together constitute the total number of scribe-created errors the percentage is lower (5.5 %). The addition and omission of verba minora is the category most often shared with other manuscripts (4 times), though this is not the largest category in the sum total of corrected readings or uncorrected singulars. Orthographic variants and harmonisations are shared three times each with other manuscripts. Twice, an omission of a word or clause is shared with other manuscripts. All these categories score relatively high in the table on page 159, where an overview is given of all the categories. There are a few marked differences between that table and the number of corrected readings shared with other manuscripts. First, as one would expect, nonsense readings are not shared. Secondly, substitutions are never shared either. Apparently, copying errors that create other words or other forms of words are not likely to be duplicated in other manuscripts, unless these can 49 Plural verbs with singular, collective substantives do frequently occur in Greek and are in themselves no evidence of Hebrew influence on the Greek text. Rahlfs adopts the singular in his text.

Singular Readings and Corrections in 1 Chronicles

163

be seen as harmonisations. Thirdly, the number of shared omissions and additions is quite high: four in the category of verba minora, one in the category of obvious improvements (the repeated en mesw twn adelfwn autwn), and two in the category of additions and omissions of words and clauses. Six out of these are shared omissions; there is only one shared addition. Final reflections on 1 Chronicles In this section we have worked from the basic assumption that each correction in the text of 1 Chronicles is a correction back towards the exemplar from which the text in Sinaiticus was copied. Given the specific circumstances of the intruded text of 1 Chronicles this is very likely but, of course, not a certainty. Though no other manuscript of 1 Chronicles has been used in the correction, it does not follow automatically that each correction is based on an actual divergence of the copied text from the exemplar. Obvious nonsense readings such as the omission of a single letter or the dittography of a syllable can be restored without reference to the exemplar. Similarly, inconsistencies in the use of pronouns show up easily when proofreading the copied text. And even in the correction phase, a scribe can make a scribal error or correct only part of an incorrectly copied reading. For example, it is problematic to assume that in 1 Chr 10:13 the exemplar read auto autw, even though the corrector has prefixed the incorrect auto before the correct autw, immediately after a line break.50 Likewise, in 1 Chr 11:2, peneis has been restored to poimeneis by inserting the letters -oim-. However, the following -ehas not been corrected to an -a- to create the correct poimaneis.51 Yet, for the large majority of the corrections, it is most likely that the exemplar was used as a basis for the corrections. If we accept the above-mentioned basic assumption, our study is relevant for evaluating the method of singular readings in a number of ways. First of all, the group of corrected readings forms an approximation of the singular readings in the text, albeit a very rough and imperfect approximation. Secondly, we are able to estimate the number of readings that arose independently in other manuscripts (between 5 and 14%). Thirdly, we have been able to demonstrate that some categories of scribecreated errors are more likely to be duplicated in other manuscripts than 50 1 Chr 10:13, folio 34.8 (OT 3), line 5.25. It is possible that we have to do here with a correction in the exemplar which was improperly understood. 51 1 Chr 11:2, folio 34.8 (OT 3), line 5.43.

164

Singular Readings and Individual Scribal Patterns

others. This is in particular the case with omissions and harmonisations. Shared substitutions—without the influence of harmonisation—are not found in this study. The intruded section of 1 Chronicles in 2 Esdras has been very helpful in forming a picture of how often corrections were made and how many singular readings were left in the text. However, it should be remembered that 1 Chronicles is a relatively poorly attested part of the Greek Bible and that it is therefore possible that a considerable number of singular readings are actually part of a wider tradition.52

SCRIBES D AND A IN PSALMS Introduction In Chapters 2 and 3 we have already studied several aspects of the book of Psalms. The book is written by the two scribes D and A, each showing a different orthographic pattern. It has been shown that the text written in red ink was added later than the normal text, and that a clear difference exists between the two scribes in the use of ligatures and nomina sacra. Possibly a difference between the scribes also exists in the choice to add a 49th line to the column to allow for a neater start of the next column. In this section we will compare the frequency and distribution of singular readings found in the work of the two scribes. The discussion below is based on the assumption that both scribes copied their text from the same exemplar. Though this point of departure cannot be proven, it is likely that this was the case. The only way in which the differences between the two scribes might be exemplar-related is if the scribes each used a different exemplar or if their exemplar consisted of two qualitatively different sections with the transition between the two sections being at the same place where one scribe took over for the other. Though possible, it is not a probable scenario. Scribe D writes the first 26.5 folios, folio 59.1 (OT 134) up to folio 62.3 recto (OT 160), leading up to LXX Ps 97:3. Scribe A, who takes over on the verso of folio 62.3 writes the remaining text up to Ps 151, which ends on folio 64.1 recto (OT 174), giving a total of 15.4 folios. As the transition occurs not only within a psalm but also neatly on the transition from the recto to verso without clear signs of stretching or condensing of 52 It could of course also be argued that the surviving manuscripts of 1 Chr are proportionally as numerous as the better attested parts of the Greek Bible. If this is the case, the “wider tradition” consists of only a limited number of manuscripts.

Scribes D and A in Psalms

165

text, we can assume that scribe A simply continued the work of the previous scribe. On the last page, the final 16 lines are left blank, and on the verso of this folio, scribe A continues with Proverbs. The data on the pricking and ruling of the quires as given in Scribes and Correctors do not show any difference that coincides with the change of hands.53 The roughening of the parchment (the pouncing) does show a difference.54 Method The text of the 151 psalms was collated against the text of Rahlfs as it has been published in the Göttingen series. After this, each of the variations was checked against the apparatus to ascertain whether it was a singular reading. As the edition of the LXX-Psalms is rather dated and almost none of the papyrus-findings of the 20th century have been included in the apparatus, a further collation of the possible singular readings of Sinaiticus was made against the text of the papyri 2149 and 2150,55 and Bodmer papyrus XXIV (Rahlfs 2110).56 The latter especially shows that quite a few presumed singular readings have a wider distribution than just Sinaiticus. Though we have more manuscripts of the Psalms than any other book of the OT in Greek, it is possible that a percentage of the singular readings in Sinaiticus are not manuscript specific but belong to a specific text-type. Our knowledge of some of the text types is limited because of the paucity of evidence. In the case of Sinaiticus, it would have been informative to have more manuscripts from what Rahlfs calls the upper and lower Egyptian types. Moreover, there is a considerable group of singular readings in Sinaiticus which Rahlfs labels as “= Masoretic Text.” Quite often these readings are not more than the removal of a single kai, which, taken by itself, can be as easily explained by transcriptional probability as by a deliberate revision towards the Hebrew. However, a sizeable number of non-singular readings are found that find additional support from the texts of Origen and the Gallican Psalter, in the latter case not infrequently through the Hexaplaric signs. Often the reading of Sinaiticus, supported by these texts, is a reading found in the Masoretic Text, increasing the Milne-Skeat, 77. See above page 35. 55 A. Pietersma, Two Manuscripts of the Greek Psalter in the Chester Beatty Library, Dublin (Rome: Biblical Institute Press, 1978). 56 R. Kasser, Papyrus Bodmer XXIV: Psaumes XVII–CXVIII (Cologny-Genève: Bibliotheca Bodmeriana, 1967). For corrections to this edition see A. Pietersma, “The Edited Text of P.Bodmer XXIV”, Bulletin of the American Society of Papyrologists 17 (1980): 67–79 53 54

166

Singular Readings and Individual Scribal Patterns

likelihood that when Sinaiticus is not supported by Origen and the Gallican Psalter but still has a Masoretic Text reading, this reading is also evidence of influence from the Masoretic Text. Because of this possible link between at least some of the “Masoretic Text” variants in Sinaiticus and a revision towards the Hebrew, we have omitted in our analysis all the variants below.57 The double line in the table separates the sections of scribe D (above) and scribe A (below the line). Rahlfs / exemplar ℵ Ps 7:7, folio 59.2 omitted kurie (OT 135), line 1.20 Ps 14:4, folio 59.3 omitted autou (OT 136), line 4.11 Ps 15:1, folio 59.3 o qeos kurie (OT 136), line 4.18 Ps 17:11, folio 59.4 ceroub ceroubin (OT 137), line 2.47 Ps 17:11, folio 59.4 kai added epetasqh (OT 137), line 2.48 Ps 17:19, folio 59.4 moi mou (OT 137), line 3.25 Ps 17:36, folio 59.4 omitted eis telos (OT 137), line 4.24 Ps 30:21, folio 59.8 anqrwpou anqrwpwn (OT 141), line 1.42 Ps 30:24, folio 59.8 omitted oti (OT 141), line 2.6 Ps 34:9, folio 60.1 en epi (OT 142), line 2.13 Ps 36:27, folio 60.2 omitted aiwnos (OT 143), line 2.4 Ps 65:5, folio 61.1 omitted kai (OT 150), line 3.7 Ps 70:8, folio 61.3 omitted opws umnhsw thn doxan (OT 152), line 2.23 sou 57 Rahlfs does not seem to be consistent in using the label “=MT.” In Ps 34:9, Sinaiticus reads en tw swthriw. The comment on the use of the preposition reads ‘= ‫’ב‬. Whether or not Rahlfs wished to indicate that this reading is an adaptation to the MT is unclear. We have classified this as a genuine singular reading and not as a Masoretic text reading.

Scribes D and A in Psalms ℵ Ps 76:16, folio 61.5 omitted (OT 154), line 2.18

Rahlfs / exemplar sou

Ps 105:48, folio 62.6 (OT 163), line 4.10 Ps 118:23, folio 63.2 (OT 167), line 1.18 Ps 120:6, folio 63.4 (OT 169), line 1.48 Ps 129:5, folio 63.5 (OT 170), line 1.26 Ps 137:3, folio 63.6 (OT 171), line 2.24

omitted

genoito

omitted

kai

kai

oude

omitted

se

omitted

se

167

One of Sinaiticus’s closest allies is Vaticanus. Unfortunately Vaticanus is incomplete for Ps 105:27–137:6. Outside this section Sinaiticus shares 44 readings with Vaticanus alone, so one would expect that part of the singular readings of Sinaiticus within the section would have been shared readings with Vaticanus had the manuscript not been defective. As the entire gap of Vaticanus falls within the text that in Sinaiticus is written by scribe A, the number of singular readings for scribe A is likely to be too high.58 In order to be able to compare the singular reading of the sections by scribes D and A, we will give not only the absolute numbers of singular readings for each category, but also the number of singular readings per 10 folios. Corrections which have been made in the same line are not always treated under the category of nonsense words but, if there is enough certainty about what was going on, under the proper category. So, cases such as autouwn (118:70) and tahn (118:36) are regarded as corrected substitutions. In the numerous tables that follow, the contributions by scribe D and scribe A are separated by a double line. Corrections by the scribes As is true for the whole poetic section of Sinaiticus (quires 59–72), the only corrections made by the early scribes in Psalms are the prima manus 58 The missing section of Vaticanus would have occupied 8 folios in Sinaiticus, which is 20% of the total text. Assuming an equal distribution of agreements one would expect 11 agreements between Sinaiticus and Vaticanus in the missing section of the latter.

168

Singular Readings and Individual Scribal Patterns

corrections; there are no indications of any distinct correction phase.59 For the whole book of Psalms the total number of corrections is 67.60 This number of corrections on 41.9 folios is markedly lower than the 90 corrections in 5 folios of 1 Chronicles. Concerning six of the corrections in Psalms, it is very difficult to be sure that they were made by the earliest scribes; a later C-corrector may have been responsible for the correction. The uncertain cases are: Ps 1:5, folio 59.1 (OT 134), line 1.16 Ps 3:2, folio 59.1 (OT 134), line 2.12 Ps 48:19, folio 60.5 (OT 146), line 3.47 Ps 71:7, folio 61.3 (OT 152), line 4.2

ℵ oi asebeis

correction asebeis

epanistantai

epanistanto

exomologhqhsetai

exomologhsetai

tou

ou

exhranqhsan Ps 101:12, folio 62.4 (OT 161), line 1.38 Ps 147:1, folio 63.8 (OT ainei [inc]s[/inc] 173), line 3.41

exhranqhn ainei

Both the uncorrected and corrected readings of Ps 3:2 are attested in manuscripts that show agreements with Sinaiticus at other places; the remaining variants are all singular readings. At eight places the original text has been erased and rewritten by the prima manus so that the initial reading has been lost.61 The erasures vary from a single letter to a series of words over two lines of text. It is remarkable that seven of these occur in the text by scribe D, and only one in that of scribe A. A clear difference in preference for this method of correction exists. I checked all these readings under UV light but nevertheless it was not always possible to determine the version of the text before the erasure. One interesting correction occurs towards the end of scribe D’s section at folio 62.3 (OT 160), line 2.1. It seems that the correction by the scribe resulted in exactly the same text as in the original As we have seen above, a seperate correction run did occur in 1 Chr, as also in the New Testament. 60 See Appendix V: ‘The Corrected Readings of Psalms’. 61 Ps 15:9, folio 59.3 (OT 136), line 4.46; Ps 77:49, folio 61.6 (OT 155), line 1.48; Ps 78:1, folio 61.6 (OT 155), line 3.31; Ps 82:18, folio 61.7 (OT 156), line 4.20; Ps 87:15, folio 61.8 (OT 157), line 4.32; Ps 93:14, folio 62.2 (OT 159), line 3.44; Ps 96:1, folio 62.3 (OT 160), line 2.1; Ps 105:10, folio 62.6 (OT 163), line 1.44. 59

169

Scribes D and A in Psalms

version, but this time neatly on the line instead of wandering off into the upper margin.62 Of the 61 certain cases of corrections by the prima manus, only two of the initial readings have additional support: (1) in Ps 87:7 eqeto occurs in the body of the Syriac text as found in the Ambrosiaster Syro-hexaplar, (2) logion instead of eleos at Ps 118:149 is found in the Psalms commentary of Hesychius of Jerusalem. In addition to these, one orthographic error was corrected by scribe A: at Ps 118:3 i mh was corrected to ei mh.63 The number of corrections by scribes D and A differ considerably when the amount of text they produced is taken into account. Scribe D has 28 corrections, which is 10.6 corrections per 10 folios; scribe A has 39 corrections which equals 25.3 corrections per 10 folios. This does not necessarily indicate that scribe A is the more precise scribe in that he corrects his text more thoroughly, but only that scribe A makes more errors of which he becomes aware while writing the text. The difference in the number of singular readings will have to decide which of the two scribes is more accurate. Discussion of the singular readings The total number of singular readings in the following analysis is 394, of which 177 belong to scribe D’s section and 217 to scribe A. The number of singular readings per 10 folios is 69.4 for scribe D and 140.9 for scribe A. And although some allowance must be made for the absence of the close allies Vaticanus and 2110 in parts of scribe A’s section, the difference remains too high not to reflect a real difference. The following table gives the numbers per category: Analytic Category

Orthography Nonsense word forms Leaps a/o verba minora

Scribe D number/10 folios (real number) 7.2 (19) 6.0 (16) 6.0 (16)64 7.2 (19)

Scribe A number/10 folios (real number) 3.9 (6) 16.2 (25) 9.1 (14)65 28.6 (44)66

This correction does not figure in the analysis of the singular readings. Ps 118:92, folio 63.2 (OT 167), line 4.45. 64 The leap in Ps 43:14 and in verse 15 is counted as one variant. 65 The leaps in Ps 118:50–51 and 130:1–2 are each counted as one variant 62 63

170

Singular Readings and Individual Scribal Patterns Analytic Category

Harmonisation - immediate context - intermediate context - general usage Editorial readings - interfering meaning - misreading syntax - misunderstood corrections - deliberate improvements Nonsense meanings Substitutions Transpositions a/o words or clauses More complex rewritings Total:

Scribe D number/10 folios (real number) 11.3 (30) 5.3 (14) 3.4 (9) 2.6 (7) 1.1 (3) 0.8 (2) 0.4 (1) 4.5 (12) 12.5 (33) 3.0 (8)67 6.0 (16) 1.9 (5) 66.8 (177)

Scribe A number/10 folios (real number) 17.5 (27) 14.9 (23) 1.9 (3) 0.6 (1) 3.9 (6) 0.6 (1) 2.6 (4) 0.6 (1) 9.7 (15) 40.3 (62) 3.9 (6) 8.4 (13) 140.9 (217)

First of all, it should be observed that the difference between scribe D and scribe A is not just a mere quantitative difference; there is a marked difference in the pattern of the singular readings. This means that scribe A is not just a “worse” scribe than scribe D, but that he is liable to some specific errors more than he is to others.

Orthography Ps 10:6, folio 59.3 (OT 136), line 1.47 Ps 16:12, folio 59.4 (OT 137), line 1.45 Ps 17:6, folio 59.4 (OT 137), line 2.31 Ps 24:15, folio 59.6 (OT 139), line 4.7

ℵ pakida[inc]s[/inc]

Rahlfs / exemplar pagidas

ws

wsei

pakides

pagides

pakidos

pagidos

66 Included here is the addition of the first moi in akouston moi poihson moi, which can feasibly be explained as misreading akouston for an imperative. 67 The three transpositions in Ps 67:32–33 are together counted as a single variant; likewise the transpositions in Ps 68:21, 71:10, and 95:7.

Scribes D and A in Psalms

Ps 30:5, folio 59.7 (OT 140), line 4.35 Ps 34:7, folio 60.1 (OT 142), line 2.8 Ps 34:8, folio 60.1 (OT 142), line 2.12 Ps 49:1, folio 60.5 (OT 146), line 4.9 Ps 54:7, folio 60.6 (OT 147), line 4.48 Ps 56:7, folio 60.7 (OT 148), line 3.31 Ps 63:6, folio 61.1 (OT 150), line 1.32 Ps 65:11, folio 61.1 (OT 150), line 3.27 Ps 67:34, folio 61.2 (OT 151), line 2.43 Ps 68:23, folio 61.2 (OT 151), line 4.33 Ps 68:31, folio 61.3 (OT 152), line 1.7 Ps 72:20, folio 61.4 (OT 153), line 2.7 Ps 77:16, folio 61.5 (OT 154), line 3.42 Ps 90:3, folio 62.2 (OT 159), line 1.15 Ps 91:4, folio 62.2 (OT 159), line 2.9 Ps 105:7, folio 62.6 (OT 163), line 1.34 Ps 108:4, folio 62.7 (OT 164), line 3.48 Ps 109:6, folio 62.8 (OT 165), line 2.16 Ps 113:4, folio 62.8 (OT 165), line 4.23 Ps 113:6, folio 62.8 (OT 165), line 4.28 Ps 118:176, folio 63.4 (OT 169), line 1.15

171

ℵ pakidos

Rahlfs / exemplar pagidos

pakidos

pagidos

pakidi

pagidi

ap anatolwn

apo anatolwn

ws

wsei

pakida

pagida

pakidas

pagidas

pakida

pagida

kat anatolas

kata anatolas

pakida

pagida

meta wdhs

met wdhs

eikonan

eikona

wsei

ws

pakidos

pagidos

meta wdhs

met wdhs

parapikranan

parepikranan

endiaballon

endieballon

krinin tois eqnesin krinei en tois eqnesin ws wsei ws

wsei

wsei

ws

172

Singular Readings and Individual Scribal Patterns

The first category comprises the orthographic variants and the number of these is higher in scribe D. These do not include any itacisms, moveable nu, or unassimilated forms such as -ng- for -gg-, which have not been noted at all.68 The 19 singular readings contain 11 instances of the change of -g- into -k- in the word pagis (spelled correctly in Ps 9:16, 31; 34:8)69 and four times a change in the way a preposition was dealt with before a vowel. apo and kata are found in their apocopated form before anatolh,70 while meta was written with the final -a- before wdh. All four examples occur in scribe D’s section. The addition of -n- to the accusative singular of the third declension is, according to Gignac, a known phenomenon. The remaining readings all involve the change of wsei to ws (Ps 77:19) or vice versa. The change between these two words also accounts for three of the six orthographic variants found in scribe A. Two variants involve the omission of the augment in compound verbs, one of which was corrected. In Ps 109:6, we find an example of inverse elision or aphaeresis, in which the first letter of the next word is elided.71

Nonsense word forms Ps 27:7, folio 59.7 (OT 140), line 2.39 Ps 38:13, folio 60.3 (OT 144), line 1.22 Ps 43:23, folio 60.4 (OT 145), line 3.1 Ps 65:8, folio 61.1 (OT 150), line 3.22 Ps 67:5, folio 61.1 (OT 150), line 4.48 Ps 68:17, folio 61.2 (OT 151), line 4.11

ℵ hapisen

Rahlfs / exemplar hlpisen

parasiwphshshs

parasiwphshs

prota

probata

ainesesews

ainesews

taracqhwsan

taracqhtwsan

oisakouson

eisakouson

The reason is that, because such orthographic variants are not mentioned in Rahlfs, comparison with other manuscripts is impossible. 69 Scribe A spells it seven times correctly with -g-. 70 See on elision Moulton-Howard, 61–2; Gignac I, 315ff; Thackeray, 136–7 ‘Elision of the final vowel of prepositions often takes place in combinations of frequent occurrence and before pronouns.’ Scribe A has apo anatwlwn at Ps 106:3 and Ps 112:3. 71 See also below, page 204. 68

Scribes D and A in Psalms

Ps 77:8, folio 61.5 (OT 154), line 3.12 Ps 77:42, folio 61.6 (OT 155), line 1.28 Ps 77:49, folio 61.6 (OT 155), line 1.48 Ps 78:1, folio 61.6 (OT 155), line 3.31 Ps 78:11, folio 61.6 (OT 155), line 4.22 Ps 80:13, folio 61.7 (OT 156), line 2.35 Ps 80:14, folio 61.7 (OT 156), line 2.40 Ps 85:14, folio 61.8 (OT 157), line 3.1 Ps 87:15, folio 61.8 (OT 157), line 4.32 Ps 93:14, folio 62.2 (OT 159), line 3.44 Ps 101:20, folio 62.4 (OT 161), line 2.10 Ps 102:15, folio 62.4 (OT 161), line 3.33 Ps 104:1, folio 62.5 (OT 162), line 2.23 Ps 104:1, folio 62.5 (OT 162), line 2.24 Ps 105:10, folio 62.6 (OT 163), line 1.44 Ps 105:38, folio 62.6 (OT 163), line 3.23 Ps 108:12, folio 62.7 (OT 164), line 4.26 Ps 117:18, folio 63.1 (OT 166), line 3.27 Ps 118:141, folio 63.3 (OT 168), line 3.25 Ps 118:148, folio 63.3 (OT 168), line 3.44 72

ℵ enwntai

Rahlfs / exemplar genwntai

ros

ceiros

[ill]erasure[/ill]len72

exapesteilen

173

klhron[ill]erasure[/ill]mi klhronomian an ka kata exapestesteila

exapesteila

isral

israhl

epanesthsthsan

epanesthsan

apwqeis thn yuchn mou [ill]erasure[/ill]kateliye ouk egkataleiyei n a[ill]erasure[/ill]u

uyou

uyous

hmerairai

hmerai

tllhlouia

allhlouia

katw

tw

[ill]erasure[/ill]twn

misountwn

efonokthnhqh

efonoktonhqh

foboum added

orfanois

kaideuwn

paideuwn

newtero

newteros

orqroun

orqron

The [ill] and [/ill] tags enclose illegible letters.

174

Singular Readings and Individual Scribal Patterns

Ps 118:174, folio 63.4 (OT 169), line 1.11 Ps 123:7, folio 63.4 (OT 169), line 3.12 Ps 126:2, folio 63.4 (OT 169), line 4.14 Ps 127:3, folio 63.4 (OT 169), line 4.33 Ps 129:8, folio 63.5 (OT 170), line 1.34 Ps 131:5, folio 63.5 (OT 170), line 2.10 Ps 134:6, folio 63.5 (OT 170), line 3.34 Ps 134:8, folio 63.5 (OT 170), line 3.38 Ps 139:2, folio 63.6 (OT 171), line 4.20 Ps 143:13, folio 63.7 (OT 172), line 4.38 Ps 143:13, folio 63.7 (OT 172), line 4.39 Ps 144:4, folio 63.8 (OT 173), line 1.11 Ps 144:13, folio 63.8 (OT 173), line 1.40 Ps 148:1, folio 63.8 (OT 173), line 4.17 Ps 150:5, folio 64.1 (OT 174), line 1.46

ℵ epepoqa

Rahlfs / exemplar epepoqhsa

wstrouqion

ws strouqion

orqizin

orqrizein

nqofuta

neofuta

pasw

paswn

w

kuriw

alassais

qalassais

prwtotoa

prwtotoka

andikou

adikou

toutoutou

toutou

plutoka

polutoka

geneaa

genea

baleia

basileia

tan

ton

euhcoiois

euhcois

A significant difference exists between the scribes in the number of nonsense words: scribe A has almost three times the number of scribe D. The nature of the errors seems to suggest that we are predominantly dealing with mere scribal slips. Of the 16 readings of scribe D, four are erased words or parts of words of which the original text could not be recovered, but which have been rewritten by the prima manus.73 A further six readings are corrected in a different manner; these readings consist mainly of an 73 Checking these rewritings under UV light did not help much. Only one of these is an erasure of one letter, Ps 78:1 klhron[ill]erasure 1 letter[/ill]mian and is quite likely just a scribal slip. The other three are all larger erasures and could be anything from sheer nonsense to harmonisation or leaps.

Scribes D and A in Psalms

175

omitted or wrong letter. Only once does this involve a syllable.74 The six uncorrected nonsense readings all consist of an omitted or doubled syllable (2 and 4 times, respectively) and once a line break is involved.75 In scribe A’s section of the text, 13 out of 25 nonsense words were corrected. The bulk of these consist of the correction of wrong letters or missing letters and syllables. The 12 uncorrected forms are also a mixture of added or reduplicated letters or syllables or an omission of one of these. There is one rewriting of an erasure, of which the initial version could not be deciphered. In this category we can isolate a number of differences between the two scribes. First of all there is the numerical difference: scribe D makes fewer errors than scribe A. Furthermore, this scribe also leaves fewer nonsense readings in the text after correction. He tends to erase text more often than scribe A so that major corrections do not disturb the optical flow of the text. Scribe A makes more errors resulting in nonsense words. He does this by substituting one letter for another or by dropping or adding a letter or a syllable somewhere in the word. Four times the nonsense word arises because the last letter of a word has been omitted.

Leaps ℵ Ps 6:9, folio 59.1 (OT omitted 134), line 4.37 Ps 9:2, folio 59.2 (OT omitted 135), line 2.40 Ps 15:9, folio 59.3 (OT [ill]erasure[/ill] 136), line 4.46 Ps 16:2, folio 59.4 (OT omitted 137), line 1.13 Ps 25:2, folio 59.6 (OT omitted 139), line 4.28 Ps 36:40, folio 60.2 omitted (OT 143), line 2.47

74 75

ros.

Rahlfs / exemplar eishkousen kurios ths fwnhs tou klauqmou mou dihghsomai panta ta qaumasia sou h sarx mou kataskhnwsei ep elpidi oi ofqalmoi mou idetwsan euquthtas dokimason me kurie kai peirason me kai bohqhsei autois kurios kai rusetai autous

ka thn corrected to kata thn; no line break involved. Ps 77:42: omission of the first syllable of ceiros. The line starts with

176

Singular Readings and Individual Scribal Patterns

ℵ Ps 43:14–15, folio 60.4 omitted (OT 145), line 2.22–23

Ps 47:14, folio 60.5 (OT 146), line 2.30 Ps 48:18, folio 60.5 (OT 146), line 3.45

omitted omitted

Ps 61:9, folio 60.8 (OT omitted 149), line 4.16 Ps 62:11, folio 61.1 (OT 150), line 1.12 Ps 67:28, folio 61.2 (OT 151), line 2.23 Ps 73:11, folio 61.4 (OT 153), line 3.20 Ps 78:6, folio 61.6 (OT 155), line 3.48

omitted omitted omitted omitted

Ps 82:18, folio 61.7 [ill]erasure[/ill]wna (OT 156), line 4.20 Ps 86:4, folio 61.8 (OT omitted 157), line 3.22 Ps 98:4, folio 62.3 (OT omitted 160), line 3.29 Ps 104:20, folio 62.5 to logion tou (OT 162), line 3.30 kuriou epurwsen auton added ou fobhqhsetai Ps 111:8, folio 62.8 (OT 165), line 3.33 etoimh h kardia autou elpizein epi ton kurion

Rahlfs / exemplar mukthrismon kai katagelwta tois kuklw hmwn eqou hmas eis parabolhn en tois eqnesin kai katadielesqe tas bareis auths oti ouk en tw apoqnhskein auton lhmyetai ta panta oude sugkatabhsetai autw h doxa autou elpisate ep auton pasa sunagwgh laou paradoqhsontai eis ceiras romfaias arcontes iouda hgemones autwn kai thn dexian sou ekceon thn orghn sou epi eqnh ta mh ginwskonta se kai taracqhtwsan eis ton aiwna outoi egenhqhsan ekei kai timh basilews krisin agapa arcwn ou mh fobhqh

Scribes D and A in Psalms

Ps 113:6, folio 62.8 (OT 165), line 4.28 Ps 117:4, folio 63.1 (OT 166), line 2.48

Ps 118:38, folio 63.2 (OT 167), line 2.7 Ps 118:50, folio 63.2 (OT 167), line 2.37 Ps 118:51, folio 63.2 (OT 167), line 2.39

Ps 118:95, folio 63.3 (OT 168), line 1.5 Ps 129:6, folio 63.5 (OT 170), line 1.29

Ps 130:1, folio 63.5 (OT 170), line 1.39 Ps 130:2, folio 63.5 (OT 170), line 1.41 Ps 133:1, folio 63.5 (OT 170), line 3.17 Ps 135:23, folio 63.6 (OT 171), line 1.26

177

eis added

Rahlfs / exemplar kai oi bounoi ws arnia probatwn eipatwsan dh pantes oi foboumenoi ton kurion oti agaqos oti eis ton aiwna to eleos autou to76

omitted

ezhsen me

ℵ omitted omitted

uperhfanoi parhnomoun ews sfodra apo de tou nomou sou eme upemeinan omitted amartwloi tou apolesai me apo fulakhs omitted prwias elpisatw israhl epi ton kurion oude eporeuqhn en omitted megalois oude eporeuqhn en alla (This reading and megalois kai the previous constitute together the swapping of two lines) en oikw kuriou omitted

omitted

omitted

oti en th tapeinwsei hmwn emnhsqh hmwn o kurios oti eis ton aiwna to eleos autou

A leap to the next line. Might also be seen as harmonisation to the immediate context, though the leap character of the variant is more obvious. 76

178

Singular Readings and Individual Scribal Patterns

Ps 138:23, folio 63.6 (OT 171), line 4.11 Ps 144:6, folio 63.8 (OT 173), line 1.20

ℵ dokimason me kai gnwqi tas tribous mou added omitted

Rahlfs / exemplar dokimason kai thn megalwsunhn sou dihgesontai

The influence of the poetic format is clearly felt in the area of leaps. Most of the leaps in this category consist of an omitted line. This may happen because of a jump from the same to the same, such as in 47:14 where two subsequent lines finish with auths and the second line is dropped. A similar circumstance arises when two lines start with the same word(s); this may also lead to one of the lines being left out (Ps 67:28). In the leap of Ps 43:14–15, the homoeoarcton involves two lines. The error made here is that the last line before the leap, which begins with eqou, was regarded as the second line of the two omitted lines, which also starts with eqou. However, often a poetic line is dropped without a similar line start or ending. Scribe D has 16 leaps, 13 of which consist of skipping one or two poetic lines (one was corrected); two consist of a downward leap of a single clause due to homoeoteleuton (one corrected),77 and in Ps 15:9 there is a leap backwards, also corrected.78 Scribe A has almost the same number of leaps (15), but in less text. We find 10 times that the scribe omitted or started to omit a whole line; in two cases the leap was noted and corrected.79 Twice we find a leap backwards (Ps 104:20 and Ps 111:8), in both cases the scribe noted and corrected his mistake. Also we find in two places a small (two and three words) corrected leap forwards, and once a leap down to the middle of the

77 The first is Ps 73:11. This verse constitutes one long poetic line in Sinaiticus. The second case is Ps 82:18, where scribe D probably jumped from -qhtwsan to -qhtwsan and only realised his mistake at the end of the verse. 78 This example has to be classified as “uncertain.” Under UV light I noticed several curves which could constitute the word ews. This could indicate a leap back to Ps 15:7. 79 Ps 130:1–2, swapping of two lines, corrected, and Ps 133:1, probably a leap to the next line.

Scribes D and A in Psalms

179

next line, which led to the creation of a whole new line.80 In total we find six corrections of leaps by scribe A. When we consider the number of dropped lines which were not corrected by the two scribes, we find that the numbers are almost the same, 5.1 per 10 folios for scribe D’s section (13 missing lines) and 4.5 for scribe A (7 missing lines). Either we have here an inherited feature of the text or both scribes were equally prone to such lapses. Though it may seem that the low number of corrections of these missing lines should favour the inherited nature of these omissions, it should not be forgotten that there is no evidence of any systematic correction in the text.

Add and omit verba minora Ps 5:10, folio 59.1 (OT 134), line 3.45 Ps 12:4, folio 59.3 (OT 136), line 2.43 Ps 18:7, folio 59.5 (OT 138), line 1.42 Ps 20:5, folio 59.5 (OT 138), line 3.23 Ps 27:2, folio 59.7 (OT 140), line 2.18 Ps 30:25, folio 59.8 (OT 141), line 2.11 Ps 36:5, folio 60.1 (OT 142), line 4.35 Ps 36:10, folio 60.2 (OT 143), line 1.4 Ps 40:10, folio 60.3 (OT 144), line 3.35 Ps 47:9, folio 60.5 (OT 146), line 2.16 Ps 56:1, folio 60.7 (OT 148), line 3.6 Ps 65:7, folio 61.1 (OT 150), line 3.14

ℵ stwmati

Rahlfs / exemplar tw stwmati

eisakouson

eisakouson mou

ouranou

tou ouranou (2)

eis ton aiwna

eis aiwna

en tw airein

en tw me airein

ton kurion

kurion

omitted

kai

omitted

kai (2)

esqiwn

o esqiwn

qeos

o qeos

daueid

tw daueid

th dunasteia

th dunasteia autou

80 Ps 138:23. The extra line consists of the first half of one stichos one and the second half of the following. Then the text follows with the unaltered first stichos. The text is not corrected so it cannot be proven that it was scribe A’s creation.

180

Singular Readings and Individual Scribal Patterns

Ps 67:36, folio 61.2 (OT 151), line 3.1 Ps 70:10, folio 61.3 (OT 152), line 2.28 Ps 77:15, folio 61.5 (OT 154), line 3.38 Ps 78:9, folio 61.6 (OT 155), line 4.10 Ps 91:15, folio 62.2 (OT 159), line 2.41 Ps 94:4, folio 62.2 (OT 159), line 4.28 Ps 96:6, folio 62.3 (OT 160), line 2.17 Ps 101:26, folio 62.4 (OT 161), line 2.27 Ps 102:7, folio 62.4 (OT 161), line 3.11 Ps 103:7, folio 62.4 (OT 161), line 4.35 Ps 104:14, folio 62.5 (OT 162), line 3.14 Ps 105:33, folio 62.6 (OT 163), line 3.10 Ps 105:34, folio 62.6 (OT 163), line 3.11 Ps 105:46, folio 62.6 (OT 163), line 4.3 Ps 107:10, folio 62.7 (OT 164), line 3.28 Ps 108:30, folio 62.8 (OT 165), line 1.40 Ps 109:3, folio 62.8 (OT 165), line 2.6 Ps 111:7, folio 62.8 (OT 165), line 3.31 Ps 113:2, folio 62.8 (OT 165), line 4.19 Ps 113:2, folio 62.8 (OT 165), line 4.20 Ps 113:23, folio 63.1 (OT 166), line 1.27

ℵ tou israhl

Rahlfs / exemplar israhl

oi ectroi mou

oi ectroi mou emoi

th erhmw

erhmw

tou onomatos

tou onomatos sou

omitted

eti

ceiri

th ceiri

omitted

kai

erga ceirwn

erga twn ceirwn

mwush

tw mwush

fwnhs sou bronths sou omitted

fwnhs bronths sou

tois ceilesin

en tois ceilesin

omitted

autois

aicmalwtisantwn oi allofuloi

twn aicmalwtisantwn allofuloi

tw stomati

en tw stomati

arch

h arch

ton kurion

kurion

h ioudaia

ioudaia

h exousia

exousia

umis estai

umeis

kai

Scribes D and A in Psalms

Ps 117:13, folio 63.1 (OT 166), line 3.18 Ps 118:1, folio 63.1 (OT 166), line 4.14 Ps 118:63, folio 63.2 (OT 167), line 3.20 Ps 118:79, folio 63.2 (OT 167), line 4.14 Ps 118:109, folio 63.3 (OT 168), line 1.37 Ps 118:115, folio 63.3 (OT 168), line 2.4 Ps 118:116, folio 63.3 (OT 168), line 2.6 Ps 118:163, folio 63.3 (OT 168), line 4.33 Ps 118:169, folio 63.3 (OT 168), line 4.47 Ps 118:173, folio 63.4 (OT 169), line 1.9 Ps 121:7, folio 63.4 (OT 169), line 2.21 Ps 123:8, folio 63.4 (OT 169), line 3.15 Ps 126:3, folio 63.4 (OT 169), line 4.19 Ps 127:4, folio 63.4 (OT 169), line 4.35 Ps 130:2, folio 63.5 (OT 170), line 1.44 Ps 133:2, folio 63.5 (OT 170), line 3.18 Ps 134:7, folio 63.5 (OT 170), line 3.35 Ps 135:11, folio 63.5 (OT 170), line 4.48 Ps 135:24, folio 63.6 (OT 171), line 1.26 Ps 138:18, folio 63.6 (OT 171), line 4.1 Ps 140:1, folio 63.7 (OT 172), line 1.24

ℵ kurios

Rahlfs / exemplar o kurios

amwmoi

oi amwmoi

fulassontwn

omitted

twn fulassontwn moi

tais cersin

tais cersin mou

181

oi ponhreuomenoi ponhreuomenoi antilabou

antilabou mou

omitted

kai

eggisatw dh

eggisatw

omitted

me

eirhnh sou

eirhnh

bohqeia

h bohqeia

ths auths gastros o anqrwpos

ths gastros anqrwpos

yuchn

thn yuchn (2)

ceiras umwn

tas ceiras umwn

ghs

ths ghs

israhl

ton israhl

ecqrwn umwn

twn ecqrwn umwn

eimi meta sou egw omitted

eimi meta sou me

182

Singular Readings and Individual Scribal Patterns

Ps 140:6, folio 63.7 (OT 172), line 1.46 Ps 142:8, folio 63.7 (OT 172), line 3.26 Ps 143:2, folio 63.7 (OT 172), line 4.4 Ps 146:1, folio 63.8 (OT 173), line 3.7 Ps 147:1, folio 63.8 (OT 173), line 3.41 Ps 147:4, folio 63.8 (OT 173), line 3.47 Ps 149:6, folio 64.1 (OT 174), line 1.19 Ps 149:6, folio 64.1 (OT 174), line 1.19 Ps 149:9, folio 64.1 (OT 174), line 1.29

ℵ rhmata mou

Rahlfs / exemplar ta rhmata mou

akouston moi

akouston

upotasswn

o upotasswn

kurion

ton kurion

ainei [inc]q[/inc] ton qn+ apostellwn

ainei ton qn+ o apostellwn

uywseis

ai uywseis

qeou

tou qeou

autois

en autois

The addition and omission of verba minora, such as articles, pronouns and conjunctions, is another area in which there is a much higher proportion of singular readings in scribe A’s section. Scribe D has 19 of these in total, and none of these has been corrected. The readings divide into 15 omissions and 4 additions. The numbers for scribe A are substantially higher: 31 omissions and 13 additions. Twice an omission was corrected.81 The figures for scribe A are almost four times as high as those for scribe D; scribe A has a higher ratio of additions than scribe D.

Harmonisations Harmonisation to the immediate context ℵ antelabeto Ps 3:6, folio 59.1 (OT 134), line 2.23 Ps 9:24, folio 59.2 epaineitai ... (OT 135), line 4.17 epi

Rahlfs / exemplar antilhmyetai To tense of previous verbs To verb: epaineitai ... epainetai en

81 Ps 118:163, omission of kai; Ps 147:1, ainei [inc]q[/inc] ton qn+. The theta is uncertain but under UV light this seemed to me the most probable reading. If true, the scribe started with the nomen sacrum initially leaving out the article. This was corrected before the mistake was completed.

183

Scribes D and A in Psalms ℵ Ps 20:7, folio 59.5 ton aiwna (OT 138), line 3.29 kai eis ton aiwna tou aiwnos Ps 45:4, folio 60.4 autwn (OT 145), line 4.45

Rahlfs / exemplar aiwna See the text of Sinaiticus in verse aiwnos 5

Ps 56:7, folio 60.7 thn yuchn (OT 148), line 3.31 Ps 64:10, folio 61.1 auths (OT 150), line 2.30

tois posin

Ps 65:20, folio 61.1 autou (OT 150), line 4.6 Ps 66:3, folio 61.1 tois laois (OT 150), line 4.15 Ps 68:12, folio 61.2 egenhqhn (OT 151), line 3.42

mou

Rahlfs: ex 4:1; more probably harmonising to autwn in previous line cf further down the verse Adapted to previous feminine singular pronouns Rahlfs: ex seq.

eqnesin

Rahlfs: ex 4-6

egenomhn

Rahlfs: cf verse 11 (passive form also) 17:19 (similar change in U R) Not in Rahlfs. Adjusted to preceeding neuter plural accusatives From second stichos of this verse auto harmonised to to agiasma in previous line Rahlfs: ex 15

Ps 77:4, folio 61.5 epoihsan (OT 154), line 2.48

autou

autwn

epoihsen

Ps 77:5, folio 61.5 eqeto enetei-lato (OT 154), line 3.2 Ps 77:69, folio 61.6 to agiasma ... to agiasma ... (OT 155), line 3.17 auto authn Ps 82:14, folio 61.7 puros (OT 156), line 4.12 Ps 93:19, folio 62.2 kardian (OT 159), line 4.10

anemou yuchn

From first half of verse

184

Singular Readings and Individual Scribal Patterns ℵ

Ps 101:12, folio 62.4 exhran(OT 161), line 1.38 qhsan

Ps 104:35, folio 62.5 (OT 162), line 4.22 Ps 105:23, folio 62.6 (OT 163), line 2.33 Ps 106:4, folio 62.6 (OT 163), line 4.23 Ps 108:14, folio 62.7 (OT 164), line 4.31 Ps 111:5, folio 62.8 (OT 165), line 3.22 Ps 112:1, folio 62.8 (OT 165), line 3.47 Ps 112:9, folio 62.8 (OT 165), line 4.14

Rahlfs / exemplar exhranqhn

corton

karpon

mh apostreyai odon polin

apostreyai

amartia

anomia

oiktmirwn

oiktirwn

autou

kuriou

odon polews

teknwn eufrainomenwn Ps 118:41, folio 63.2 eleos (OT 167), line 2.14

teknwn eufrainomenhn logion

Ps 118:69, folio 63.2 sou (OT 167), line 3.36

mou

Ps 118:88, folio 63.2 sou (OT 167), line 4.37

tou

Ps 118:124, folio 63.3 (OT 168), line 2.33 Ps 118:134, folio 63.3 (OT 168), line 3.8 Ps 118:147, folio 63.3 (OT 168), line 3.40 Ps 124:4, folio 63.4 (OT 169), line 3.30

kata ta ta dikaiwdikaiw-mata mata fulaxai

fulaxw

proefqa-san proefqasa me euqes added

Not in Rahlfs. Grammatically a first person is required. Previous verb, also end of line, -qhsan Rahlfs: ex first line Rahlfs: ex seq.

agaqois

Rahlfs: ex first line Rahlfs: ex 4 Harmonisation to next phrase.

Rahlfs: ex previous line cf vss 116, 124, 149 Influence of earlier occurrence in line Not in Rahlfs. sou three words later Rahlfs: ex previous line repeated Not in Rahlfs. Influenced by previous verb Verb from next verse. Pronoun to make sense From further down in verse

Scribes D and A in Psalms ℵ Ps 139:12, folio 63.7 mh (OT 172), line 1.9 kateuqunqhsetai Ps 139:14, folio 63.7 tw (OT 172), line 1.19 proswpou sou Ps 140:1, folio 63.7 ths fwnhs (OT 172), line 1.22

Ps 140:7, folio 63.7 (OT 172), line 2.2

Ps 144:17, folio 63.8 (OT 173), line 2.11 Ps 144:20, folio 63.8 (OT 173), line 2.20 Ps 148:10, folio 63.8 (OT 173), line 4.39 Ps 149:9, folio 64.1 (OT 174), line 1.31

185

Rahlfs / exemplar Rahlfs: ex 11 kateuqunqhsetai tw proswpw Not in Rahlfs sou th fwnh

Rahlfs: adapted to first stichos? Probably not, more likely influence of next word dieskorpidieskorpi-sqh Rahlfs: cf 11:3. Rahlfs is unlikely sqhsan to be correct here. Probably the plural is continuing the earlier plural subject osiois osios Many dative plurals around exole-qreuei exole-qreusei Rahlfs: ad fulassei adapt erpetina erpeta Influence of petina autwn autou Adaption to en autois in previous line

The number of harmonisations to the immediate context shows a similar pattern to that seen before: Scribe D has considerably fewer of these singular readings than scribe A. Scribe D corrects only one, against five corrections by scribe A. Even after subtracting the corrections by the first hand still twice as many errors of this type remain in scribe A’s section. It should be observed that it is not uncommon for the source of harmonisation to come from further down in the text instead of from a passage just written. So we find, for example, in Ps 44:10 pepoikilmenois from verse 14 (in the text of Sinaiticus); thn yuchn for tois posin from the next line; Ps 65:20 autou for mou from the next clause (note that the whole verse is one stichos in

186

Singular Readings and Individual Scribal Patterns

Sinaiticus instead of two); Ps 66:3 tois laois for eqnesin as in the next two lines. Harmonisation to the intermediate context Ps 44:10, folio 60.4 (OT 145), line 4.3

Ps 51:11, folio 60.6 (OT 147), line 3.25 Ps 66:5, folio 61.1 (OT 150), line 4.20

Ps 68:35, folio 61.3 (OT 152), line 1.18

Ps 70:2, folio 61.3 (OT 152), line 2.6 Ps 77:40, folio 61.6 (OT 155), line 1.23 Ps 77:48, folio 61.6 (OT 155), line 1.45 Ps 88:4, folio 62.1 (OT 158), line 1.6

Rahlfs / exemplar pepoikil-menh Rahlfs: cf v.14 (correct, in v.14 it is grammatically plausible, not here in v.10) ephkousas epoihsas Rahlfs: ex 117:21, 28 mou krinei thn krineis Rahlfs: ex 9:9, 95:13, 97:9. Next oikoume-nhn line is also 9:9, en definitely influence dikaiosunh from parallel added passages in Psalms perata ths erponta en Rahlfs: ex 66:8. Triggered by the ghs autois context and confusing the consonants of erponta kai exelou me Rahlfs: cf 30:2 omitted Previous verse also changed towards 30:2 th erhmw gh anudrw Rahlfs: ex 40:1. Possibly same variation in 2110 calazan aicmalwRahlfs: ex 61 sian wmosa uywsa Rahlfs: ex 20. There is no eklekton obvious trigger for ek tou laou the influence mou added except the verb dieqemhn ℵ pepoikilmenois

187

Scribes D and A in Psalms

Ps 96:9, folio 62.3 (OT 160), line 2.29

ℵ epi pantas tous qeous

Rahlfs / exemplar uper pantas Rahlfs: ex 1 or 94:3, 95:4. Equally tous qeous possible is the inflence of uper in the second stichos of this verse

Ps 108:26, folio 62.8 mega eleos (OT 165), line 1.27 Ps 118:60, folio 63.2 fulaxas(OT 167), line 3.11 qai

eleos

Rahlfs: ex 50:3

tou fulaxasqai

Ps 131:7, folio 63.5 (OT 170), line 2.15

eiseleusomeqa

Rahlfs: cf verse 57. Infinitive in v. 57 also without tou Not simply a dropped syllable but harmonisation to vss 3–5

eiseleusome

When it comes to harmonisations to the intermediate context, we fail to find the familiar pattern. On the available evidence it even seems that scribe D displays a stronger inclination to this error than scribe A. But the numbers involved are relatively small and need to be treated with caution. Harmonisation to remote context / general usage Ps 20:8, folio 59.5 (OT 138), line 3.36

ℵ saleuqw

Ps 29:1, folio 59.7 (OT 140), line 3.34

yalmos t

Ps 40:12, folio 60.3 (OT 144), line 3.39

hqelhsas

Ps 41:6, folio 60.3 (OT 144), line 4.23

swthrion tou proswpou sou daueid

Ps 78:1, folio 61.6 (OT 155), line 3.30

Rahlfs / exemplar saleuqh Rahlfs notes other places of ou mh saleuqw yalmos Most probably intended yalmos tw da+d teqelhkas Change to aorist: this would be the only occurrence of the perfect of qelw swthrion tou proswpou mou asaf

188

Singular Readings and Individual Scribal Patterns

Ps 90:12, folio 62.2 (OT 159), line 1.40 Ps 96:7, folio 62.3 (OT 160), line 2.18

ℵ kai epi ceirwn pepoiqo-tes epi

Ps 118:19, folio 63.2 apostre(OT 167), line 1.8 yhs

Rahlfs / exemplar epi ceirwn Rahlfs: ex Matt 4:6 (possibly) proskuRahlfs: ex Isa 42:17 (probably) nountes apokruyhs

apostrefein often occurs in Psalms

A few readings are listed as harmonisations to remote contexts or to general usage. The change of the perfect teqelhkas to the aorist hqelhsas removes the only perfect form of this verb in the whole Greek Bible. The phrase swthrion tou proswpou mou was changed to proswpon tou proswpou sou, and the verb apokruyhs was replaced with apostreyhs. Scribe D shows a stronger tendency to this type of harmonisation than scribe A. Scribe D has also two readings that reflect influence from remote passages. This influence, from Matthew and Isaiah, points to an intimate knowledge of the biblical text.

Editorial readings Ps 45:4, folio 60.4 (OT 145), line 4.43 Ps 73:4, folio 61.4 (OT 153), line 2.46

ℵ kumata orghs

Ps 103:12, folio pterwn 62.5 (OT 162), line 1.2 Ps 106:14, folio en skotous 62.7 (OT 164), kai ek skias line 1.8

Rahlfs / exemplar udata Interfering meaning: substitution with associated word eorths Interfering meaning: line speaks about judgment petrwn ek skotous kai skias

Ps 118:75, folio etapinw-san etapeinw63.2 (OT 167), sas line 4.4

Interfering meaning: subject is the birds of heaven Misreading syntax: missed the parallelism of the two nouns Misreading syntax: all the preceeding nouns taken as subject of the verb

Scribes D and A in Psalms

189

Rahlfs / exemplar ℵ Ps 118:82, folio parekalesan pote Misreading syntax: 63.2 (OT 167), parakaleseis Whole verse one line 4.24 stichos in ℵ, previous subject retained Ps 118:73, folio eplasan kai epoihsan kai Misunderstood 63.2 (OT 167), me htoimasan me eplasan correction: see main line 3.44 text Ps 40:8, folio omitted kat emou (2) Deliberate 60.3 (OT 144), improvement: line 3.28 removal of repeated phrase

There does not seem to be much difference between the two scribes in most subcategories of the more editorially coloured singular readings except for one group, namely, that of misinterpreting syntax. All the examples are by scribe A. The three examples of interfering meaning (two by scribe D and one by scribe A) all have in common that the actual physical shape of the replacing word resembles that of the replaced word. In Ps 118:73, we find a reading that may have arisen out of a misunderstood correction in the exemplar. Sinaiticus reads the verbs eplasan ... htoimasan, the two known readings are epoihsan ... eplasan and epoihsan ... htoimasan. It is possible that Sinaiticus’s somewhat conflated reading is the result of a misinterpretation of a correction.

Nonsense meanings Ps 31:8, folio 59.8 (OT 141), line 2.44 Ps 44:17, folio 60.4 (OT 145), line 4.25 Ps 45:6, folio 60.5 (OT 146), line 1.3 Ps 56:2, folio 60.7 (OT 148), line 3.10 Ps 56:4, folio 60.7 (OT 148), line 3.18 Ps 60:6, folio 60.8 (OT 149), line 3.28 Ps 68:18, folio 61.2 (OT 151), line 4.17

ℵ omitted

Rahlfs / exemplar h (dative relative pronoun)

egenhqh

egenhqhsan

bohqhsai

bohqhsei

h yuchn

h yuch

tou

tous

o

to

ta cu+

tacu

190

Singular Readings and Individual Scribal Patterns

Ps 71:7, folio 61.3 (OT 152), line 4.2 Ps 76:5, folio 61.5 (OT 154), line 1.40 Ps 80:6, folio 61.7 (OT 156), line 2.13 Ps 84:6, folio 61.8 (OT 157), line 1.34 Ps 85:9, folio 61.8 (OT 157), line 2.33 Ps 100:3, folio 62.3 (OT 160), line 4.26 Ps 104:39, folio 62.5 (OT 162), line 4.35 Ps 105:2, folio 62.6 (OT 163), line 1.14 Ps 113:1, folio 62.8 (OT 165), line 4.17 Ps 118:102, folio 63.3 (OT 168), line 1.21 Ps 118:112, folio 63.3 (OT 168), line 1.47 Ps 118:127, folio 63.3 (OT 168), line 2.38 Ps 120:8, folio 63.4 (OT 169), line 2.3 Ps 121:7, folio 63.4 (OT 169), line 2.23 Ps 132:2, folio 63.5 (OT 170), line 3.8 Ps 137:7, folio 63.6 (OT 171), line 2.37 Ps 138:5, folio 63.6 (OT 171), line 3.6 Ps 138:6, folio 63.6 (OT 171), line 3.11 Ps 138:19, folio 63.6 (OT 171), line 4.2 Ps 147:5, folio 63.8 (OT 173), line 4.2

ℵ tou

Rahlfs / exemplar ou

o pantes oi ecqroi

oi ofqalmoi

exelqen

exelqein

genean

geneas

proskunhsou

proskunhsousin

poiounta para basileis nefelh

poiountas parabaseis nefelhn

akousta

akoustas

oikoi

oikou

klimatwn

krimatwn

dia pantos amiyein to

di antameiyin sou

fulaxeis

fulaxei

tois purgobaresin

tais purgobaresin

katabainwn

katabainon

exetina ceiras so[inc]u[/inc] dikaia

exeteinas ceiras sou arcaia

auton

authn

ex amartwlous

amartwlous

omiclh

omiclhn

Scribes D and A in Psalms

191

There is a relative difference between the scribes in the number of nonsense meanings, a category which is closely connected with the pure nonsense reading. The 12 readings by scribe D look like scribal slips with the difference that the result is an existing word: an article o before pantes, the omission of a final syllable (proskunhsou for proskunhsousin), and the article tou instead of the required tous. Interesting is the confusion the word tacu caused in Ps 68:18; the last two letters were interpreted as the nomen sacrum for cristou. Five out of twelve readings have been corrected. The 15 readings of scribe A may have a slightly higher component of visual errors: para basileis for parabaseis and dikaia for arcaia. We also see that the final letter of a word is omitted five times. Out of the fifteen readings, three were corrected by the scribe. Though the main difference between the two scribes seems to be one of quantity, the dropping of single letters at the end of words by scribe A is noteworthy.

Substitutions Ps 15:6, folio 59.3 (OT 136), line 4.34 Ps 19:8, folio 59.5 (OT 138), line 3.2 Ps 21:15, folio 59.5 (OT 138), line 4.45 Ps 22:1, folio 59.6 (OT 139), line 2.13 Ps 30:21, folio 59.8 (OT 141), line 1.40 Ps 33:7, folio 59.8 (OT 141), line 4.38 Ps 34:9, folio 60.1 (OT 142), line 2.13 Ps 34:27, folio 60.1 (OT 142), line 3.23 Ps 36:8, folio 60.1 (OT 142), line 4.45 Ps 36:15, folio 60.2 (OT 143), line 1.17 Ps 43:17, folio 60.4 (OT 145), line 2.30

ℵ epesan

Rahlfs / exemplar epepesan

agalliasomeqa

megalunqhso-meqa

diabhmata

osta

usterhsh

usterhsei

apokrufois

apokrufw

autwn

autou

en tw swthriw

epi tw swthriw

qeos

kurios

en tw ponhreuesqai yuchn

wste ponhreuesqai thn kardian

fobou

proswpou

192

Singular Readings and Individual Scribal Patterns

Ps 48:15, folio 60.5 (OT 146), line 3.36 Ps 54:10, folio 60.7 (OT 148), line 1.10 Ps 55:5, folio 60.7 (OT 148), line 2.25 Ps 64:5, folio 61.1 (OT 150), line 2.11 Ps 64:9, folio 61.1 (OT 150), line 2.24 Ps 64:10, folio 61.1 (OT 150), line 2.29 Ps 67:22, folio 61.2 (OT 151), line 2.8 Ps 67:31, folio 61.2 (OT 151), line 2.32 Ps 68:7, folio 61.2 (OT 151), line 3.27 Ps 68:20, folio 61.2 (OT 151), line 4.21 Ps 69:2, folio 61.3 (OT 152), line 1.31 Ps 73:6, folio 61.4 (OT 153), line 3.2 Ps 73:13, folio 61.4 (OT 153), line 3.24 Ps 78:6, folio 61.6 (OT 155), line 3.49 Ps 78:9, folio 61.6 (OT 155), line 4.13 Ps 81:7, folio 61.7 (OT 156), line 3.20 Ps 82:6, folio 61.7 (OT 156), line 3.35 Ps 84:14, folio 61.8 (OT 157), line 2.9 Ps 85:1, folio 61.8 (OT 157), line 2.12 Ps 85:14, folio 61.8 (OT 157), line 3.3 Ps 88:35, folio 62.1 (OT 158), line 3.2

ℵ epalaiwqh

Rahlfs / exemplar palaiwqhsetai

gh

polei

kuriw

qew

katoikhsei

kataskhnwsei

thn ghn

ta perata

udatw

udatwn

plhmmelia

plhmmeleiais

[ill]erasure[/ill]

lawn

ekzhtountes

zhtountes

emou

mou

qelhson

speuson

diekoyan

exekoyan

dunasteia

dunamei

eidota

ginwskonta

autou

sou

arcwn

arcontwn

kai

oti

poreusetai

proporeusetai

tou daueid

tw daueid

krataioi

sunagwgh krataiwn thn diaqhkhn

en th diaqhkh

Scribes D and A in Psalms ℵ Ps 94:6, folio 62.2 (OT enwpion 159), line 4.36 Ps 100:6, folio 62.3 (OT 160), line 4.34 Ps 102:2, folio 62.4 (OT 161), line 2.45 Ps 102:15, folio 62.4 (OT 161), line 3.35 Ps 103:3, folio 62.4 (OT 161), line 4.23 Ps 103:8, folio 62.4 (OT 161), line 4.39 Ps 103:22, folio 62.5 (OT 162), line 1.28 Ps 103:28, folio 62.5 (OT 162), line 1.46 Ps 104:16, folio 62.5 (OT 162), line 3.19 Ps 105:21, folio 62.6 (OT 163), line 2.26 Ps 106:30, folio 62.7 (OT 164), line 2.13 Ps 107:3, folio 62.7 (OT 164), line 3.5 Ps 107:12, folio 62.7 (OT 164), line 3.32 Ps 108:19, folio 62.8 (OT 165), line 1.5 Ps 108:21, folio 62.8 (OT 165), line 1.11 Ps 109:6, folio 62.8 (OT 165), line 2.18 Ps 110:6, folio 62.8 (OT 165), line 2.39 Ps 110:6, folio 62.8 (OT 165), line 2.40 Ps 111:1, folio 62.8 (OT 165), line 3.13 Ps 113:11, folio 62.8 (OT 165), line 4.40 Ps 114:1, folio 63.1 (OT 166), line 1.40

193

Rahlfs / exemplar enantion

autou

mou

apodosis

antapodoseis

oti ei

wsei

nefelhn

nefh thn

autous

autois

epi tais mandrais

en tais mandrais

panta

sumpanta

autou

artou

kuriou

tou qeou

epimelia

epi limena

egerqhti

exegerqhti

ouci

ouk

zwnnutai

perizwnnutai

eneka

eneken

ghn pollhn

ghs pollwn

eautou

autou

eautou

autou

qelei

qelhsei

epi ths ghs

en th gh

o qeos

kurios

194

Singular Readings and Individual Scribal Patterns

Ps 114:2, folio 63.1 (OT 166), line 1.43 Ps 117:5, folio 63.1 (OT 166), line 3.1 Ps 118:15, folio 63.1 (OT 166), line 4.48 Ps 118:34, folio 63.2 (OT 167), line 1.45 Ps 118:35, folio 63.2 (OT 167), line 2.1 Ps 118:36, folio 63.2 (OT 167), line 2.2 Ps 118:49, folio 63.2 (OT 167), line 2.35 Ps 118:55, folio 63.2 (OT 167), line 2.47 Ps 118:70, folio 63.2 (OT 167), line 3.38 Ps 118:72, folio 63.2 (OT 167), line 3.42 Ps 118:77, folio 63.2 (OT 167), line 4.10 Ps 118:109, folio 63.3 (OT 168), line 1.38 Ps 118:114, folio 63.3 (OT 168), line 2.3 Ps 118:114, folio 63.3 (OT 168), line 2.3 Ps 118:122, folio 63.3 (OT 168), line 2.25 Ps 120:4, folio 63.4 (OT 169), line 1.43 Ps 120:6, folio 63.4 (OT 169), line 1.47 Ps 122:3, folio 63.4 (OT 169), line 2.40 Ps 122:4, folio 63.4 (OT 169), line 2.42 Ps 127:6, folio 63.4 (OT 169), line 4.40 Ps 130:2, folio 63.5 (OT 170), line 1.45

ℵ autou

Rahlfs / exemplar mou

moi

mou

ekzhthsw

katanohsw

ekzhthsw

exereunhsw

auton

authn

tahn

thn

twn doulwn

tw doulw

to onoma

tou onomatos

autouwn

autwn

mou

sou

moi

mou

ton nomon

tou nomou

laon

logon

hlpisa

ephlpisa

agaqa

agaqon

exupnwsei

upnwsei

ouk ekkausei

ou sugkausei

eplhqunqhmen

eplhsqhmen

eplhqunqh

eplhsqh

idhs

idois

ews

ws

Scribes D and A in Psalms

Ps 131:13, folio 63.5 (OT 170), line 2.35 Ps 134:10, folio 63.5 (OT 170), line 3.43 Ps 134:18, folio 63.5 (OT 170), line 4.17 Ps 135:14, folio 63.6 (OT 171), line 1.7 Ps 136:3, folio 63.6 (OT 171), line 1.39 Ps 136:3, folio 63.6 (OT 171), line 1.41 Ps 138:9, folio 63.6 (OT 171), line 3.16 Ps 140:4, folio 63.7 (OT 172), line 1.36 Ps 142:2, folio 63.7 (OT 172), line 3.2 Ps 143:2, folio 63.7 (OT 172), line 4.4 Ps 143:3, folio 63.7 (OT 172), line 4.5 Ps 143:15, folio 63.7 (OT 172), line 4.45 Ps 143:15, folio 63.7 (OT 172), line 4.46 Ps 144:1, folio 63.7 (OT 172), line 4.48 Ps 144:1, folio 63.7 (OT 172), line 4.48 Ps 144:10, folio 63.8 (OT 173), line 1.30 Ps 144:19, folio 63.8 (OT 173), line 2.16 Ps 145:4, folio 63.8 (OT 173), line 2.34 Ps 145:5, folio 63.8 (OT 173), line 2.37 Ps 147:4, folio 63.8 (OT 173), line 3.48 Ps 147:7, folio 63.8 (OT 173), line 4.9

ℵ exelaxato

Rahlfs / exemplar exelexato

apektinen

os epataxen

en autois

ep autois

en mesw

dia mesou

hrwthsan

ephrwthsan

umnhsate

umnon asate

laboimi

analaboimi

enduasw

sunduasw

twn doulwn

tou doulou

sou

mou

egnwrisqhs

egnwsqhs

emakarisa

emakarisan

sou

ou

qee

o qeos

basileu

o basileus

exomologeisqwsan epakousei

exomologhsasqwsan epakousetai

autos

autou (2)

sou

ou

ws tacus

ews tacous

pneusai

pneusei

195

196

Singular Readings and Individual Scribal Patterns

In the category of substitutions, a clear difference between the two scribes appears. The totals (33 for scribe D and 62 for scribe A) can be broken down further by the nature of the substitution. The majority of the substitutions concern the substitution by another word, which can mean a different personal pronoun (change in person), or a different preposition or noun. Differences involving case and/or number of (pro)nouns constitute the second subdivision. The treatment of prefixed prepositions forms the third group, and the final group is formed by the substitution of one form of the verb by another. The numbers break down as follows: Type of substitution

Total

Different word Case/number82 Prefixed prepositions total: - omitted - added - substituted Verb forms

49 23 15 9 2 4 8

Scribe D number/10 folios (real number) 7.2 (19) 3.0 (8) 1.5 (4) 0.8 (2) 0.4 (1) 0.4 (1) 0.8 (2)

Scribe A number/10 folios (real number) 19.5 (30) 9.7 (15) 7.1 (11) 4.5 (7) 0.6 (1) 1.9 (3) 3.9 (6)

One cannot say much about the differences between the scribes when the number of variants is low, but some differences stand out clearly. Scribe A has the tendency to replace a word (but corrects himself twice) and is also less careful with case/number substitutions (again, two corrections by scribe A and one by D) and verb forms. A marked characteristic in his handling of prefixed prepositions is that scribe A tends to drop these rather than add.

Transpositions Ps 44:18, folio 60.4 (OT 145), line 4.29 Ps 67:32, folio 61.2 (OT 151), line 2.38 Ps 67:33, folio 61.2 (OT 151), line 2.39 Ps 67:33, folio 61.2 (OT 151), line 2.40

ℵ sou tou onomatos

Rahlfs / exemplar tou onomatos sou

kuriw

qew

kuriw

qew

qew

kuriw

Note that the changes in case/number that can be attributed to the influence of the immediate context have not been included here. 82

Scribes D and A in Psalms

Ps 68:21, folio 61.2 (OT 151), line 4.28 Ps 68:21, folio 61.2 (OT 151), line 4.29 Ps 71:10, folio 61.3 (OT 152), line 4.11 Ps 71:10, folio 61.3 (OT 152), line 4.12 Ps 72:6, folio 61.4 (OT 153), line 1.17 Ps 75:12, folio 61.5 (OT 154), line 1.20 Ps 87:16, folio 61.8 (OT 157), line 4.35 Ps 95:7, folio 62.3 (OT 160), line 1.24 Ps 95:7, folio 62.3 (OT 160), line 1.25

ℵ euron

Rahlfs / exemplar uphrxen

ouk uphrcen

ouc euron

prosaxousin

prosoisousin

prosoiousin

prosaxousin

197

asebeian kai adikian adikian kai asebeian tw kuriw qew kuriw tw qew en kopois kai

kai en kopois

enegkate doxan kai timhn

enegkate ai patriai twn eqnwn enegkate doxan kai timhn

enegkate ai patriai twn eqnwn

wsei diploida aiscunhn ta krimata sou kai alhqeia erwthsate dh ta eis eirhnhn thn ierousalhm Ps 121:6, folio 63.4 erwthsate dh ta eis kai euqhnia tois (OT 169), line 2.20 eirhnhn thn agapwsin se ierousalhm megalh h doxa Ps 137:5, folio 63.6 h doxa kuriou (OT 171), line 2.32 megalh kuriou Ps 139:11, folio 63.7 katabaleis autous epi ths ghs (OT 172), line 1.7 epi ths ghs katabaleis autous auth estin Ps 149:9, folio 64.1 estin auth (OT 174), line 1.30

Ps 108:29, folio 62.8 (OT 165), line 1.37 Ps 118:75, folio 63.2 (OT 167), line 4.3 Ps 121:6, folio 63.4 (OT 169), line 2.19

aiscunhn wsei diploidan kai alhqia ta krimata sou kai euqhnia tois agapwsin se added

There is not much difference in the number of transpositions, though the relative number in scribe A’s section is marginally higher.83 In the texts In Ps 130:1–2 two lines have been swapped by scribe A as the result of a jump down. We have counted that variant under leaps. 83

198

Singular Readings and Individual Scribal Patterns

of both scribes D and A, a transposition of two poetic lines occurs (at Ps 95:7 and 121:6). Scribe D has four instances where two clauses within a single line are swapped, scribe A has five. In Ps 67:32–33 confusion arises regarding qw+ and kw+. A transposition of the verb of one line with the verb in the next line occurs at Ps 68:21 and 71:10. This type of transposition does not occur in scribe A’s section.

Addition and omission of words and phrases Ps 27:3, folio 59.7 (OT 140), line 2.21 Ps 30:23, folio 59.8 (OT 141), line 1.48 Ps 58:2, folio 60.7 (OT 148), line 4.46 Ps 61:7, folio 60.8 (OT 149), line 4.11 Ps 67:17, folio 61.2 (OT 151), line 1.37 Ps 67:34, folio 61.2 (OT 151), line 2.44 Ps 68:32, folio 61.3 (OT 152), line 1.10 Ps 73:10, folio 61.4 (OT 153), line 3.17 Ps 76:9, folio 61.5 (OT 154), line 2.1 Ps 76:14, folio 61.5 (OT 154), line 2.15 Ps 77:35, folio 61.6 (OT 155), line 1.6 Ps 83:4, folio 61.7 (OT 156), line 4.34 Ps 85:4, folio 61.8 (OT 157), line 2.21 Ps 87:1, folio 61.8 (OT 157), line 3.36 Ps 89:2, folio 62.1 (OT 158), line 4.7 Ps 96:5, folio 62.3 (OT 160), line 2.14

ℵ andrwn ergazomenwn omitted

Rahlfs / exemplar ergazomenwn

ke+ ek twn ecqrwn

ek twn ecqrwn

omitted

mh

omitted

o qeos

omitted

idou

omitted

ekferonta

omitted

o qeos

omitted

apokoyei

omitted

qeos

omitted

o qeos

omitted

eauth

mou o qeos

mou

omitted

eis to telos

omitted

su ei

omitted

kuriou

proswpou

199

Scribes D and A in Psalms

Ps 101:16, folio 62.4 (OT 161), line 1.48 Ps 114:5, folio 63.1 (OT 166), line 2.2 Ps 115:3, folio 63.1 (OT 166), line 2.20 Ps 118:20, folio 63.2 (OT 167), line 1.11 Ps 118:56, folio 63.2 (OT 167), line 3.1 Ps 118:64, folio 63.2 (OT 167), line 3.22 Ps 118:152, folio 63.3 (OT 168), line 4.5 Ps 118:169, folio 63.3 (OT 168), line 4.48 Ps 121:4, folio 63.4 (OT 169), line 2.14 Ps 121:9, folio 63.4 (OT 169), line 2.28 Ps 139:5, folio 63.6 (OT 171), line 4.29 Ps 143:11, folio 63.7 (OT 172), line 4.29 Ps 149:1, folio 64.1 (OT 174), line 1.4

ℵ

Rahlfs / exemplar

omitted

ta eqnh

kurios o qeos

o qeos

omitted

pantwn

eis ta krimata

ta krimata

omitted

oti

pasa h gh

h gh

omitted

egnwn

kurie kurie (verse one stichos) omitted

kurie

omitted

agaqa

omitted

kurie

omitted

exelou me

omitted

allhlouia

fulai

When it comes to the omission and addition of words and phrases that cannot be explained by any of the previous categories, there is again not much difference between the two scribes, albeit that scribe A, again, has the higher number. Among the 16 singular readings by scribe D are seven variants that involve qeos or kurios (six omissions; one addition). In scribe A’s 13 readings, the same thing happens three times (one omission; two additions).

Major rewritings ℵ Ps 36:23, folio 60.2 (OT 143), kateuqunetai line 1.41 ta diabhmata andri

Rahlfs / exemplar ta diabhmata anqrwpou kateuqunetai

200

Singular Readings and Individual Scribal Patterns

Ps 54:21, folio 60.7 (OT 148), line 1.43 Ps 54:21, folio 60.7 (OT 148), line 1.43 Ps 54:22, folio 60.7 (OT 148), line 1.45 Ps 65:13, folio 61.1 (OT 150), line 3.35 Ps 68:25, folio 61.2 (OT 151), line 4.40 Ps 71:19, folio 61.3 (OT 152), line 4.46

ℵ omitted omitted kai added tas eucas mou apodwsw sou laboi

omitted

Ps 71:19, folio 61.3 (OT 152), en pash th gh line 4.46

Rahlfs / exemplar exeteinen thn ceira autou ebebhlwsan thn diaqhkhn autou diemerisqhsan apodwsw soi tas eucas mou ths orghs sou katalaboi eis ton aiwna kai eis ton aiwna tou aiwnos kai plhrwqhsetai ths doxhs autou pasa h gh

Major rewritings, those in which more than one process is going on, occur exclusively in scribe D’s section (five variant clusters). In Ps 36:23 we find a combination of a transposition and a substitution: andri instead of anqrwpou. In Ps 54:21 and 71:19, a large omission occurs with some repairs made immediately after the omitted text in order to maintain the flow of the text. Conclusions When the work of the two scribes is compared with each other, the general assessment must be that scribe A produces many more singular readings than scribe D. However, our study has enabled us to go beyond a mere general assessment. Leaving aside the orthographic variants such as those having to do with ei–i and ai–e, moveable nu, and assimilation of two consonants, we notice that scribe D’s text shows more remarkable orthographic features than scribe A’s does. The main inconsistency lies in the spelling of pagis, while the treatment of prepositions ending in a vowel is not always consistent. An interesting insight into scribe D’s copying technique may be the observation that his nonsense readings are not only infrequent and mostly corrected, but also that the unnoticed mistakes often concern repeated syllables.

Scribes D and A in Psalms

201

Scribe A’s nonsense readings betray the imprecision with which he works. Also, he is more inclined to drop a prefixed preposition than scribe D, and when he accidentally writes the wrong letter he does not take the time to expunge and rewrite the text but simply places a dot above the letter and continues with the correct one. Harmonisations to the immediate context are also more numerous with this scribe. Though the text by scribe A has a high number of ei–i and ai–e interchanges, which is an aural rather than an optic feature, the higher frequency of errors that are influenced by a similar shape of the letters is a marked difference with scribe D. Leaps backward were easier to detect than omissions that arose because of a leap forward, especially when the leap back stretches over one or more lines. Therefore, such leaps were often corrected. Scribe A twice corrects a leap forward. One wonders whether he copies Psalms per poetic line, as this would be in line with the high number of changes to verba minora. Changes in this category are more likely to happen when larger chunks of text have to be retained in the short-term memory than in text of only a few words length. Scribe D is influenced more by the intermediate context and remote passages than scribe A. Also, all the major rewritings can be found in his section. Scribe A misinterprets at times the syntax of a passage, leading to a rewriting of one or two clauses. This type of error does not occur with scribe D. The frequent occurrence of missing poetic lines indicates that the exemplar, or an ancestor further down if the missing lines are an inherited feature, was written in a format in which each poetic line was attributed its own physical line (as in Sinaiticus) and unlike, for example, the Bodmer papyrus XXIV (Rahlfs 2110), where poetic lines are only marked off by interpunction. The book of Psalms stands out among the other books of the Old Testament (with the exception of Song of Songs) by the use of red ink for the headings and numbers. Though this apparent care for the appearance of Psalms is obviously a concern for the makers of Sinaiticus, this same care is not extended to the actual text. No further checking of the text takes place after it is first written down, and the result is that some obvious errors, such as the missing heading of Ps 149, remain unnoticed.

202

Singular Readings and Individual Scribal Patterns

SCRIBES D AND A IN PAUL Introduction Our third sample of singular readings covers the readings of five books from the Pauline corpus: Romans, Colossians, 1 and 2 Thessalonians, and Hebrews. Romans follows the Gospel of John on a new quire and covers 6.64 folios. Colossians, 1 and 2 Thessalonians, and Hebrews form one block of 9.77 folios.84 Most of this text is written by scribe A; folios 84.3 and 84.6 (NT 88 and 91) form a replacement sheet written by scribe D. They contain respectively the last part of 1 Thessalonians (from 2:14 onwards) and Heb 4:16–8:1. The total text written by scribe A covers 14.49 folios, the text by scribe D covers 1.92 folio. This sample was chosen because of the presence of the two replacement folios by scribe D in the middle scribe A’s work, which gives us a further opportunity to compare the characteristics of these two scribes. In the main table the figures are adjusted to the number of errors per 10 folios. Especially for scribe D’s sample, one should not attach too much significance to these numbers as they are derived from a sample of less than two folios. The singular status of the readings was established using the apparatuses of NA27, von Soden, and Tischendorf’s Editio Octava Critica Maior.85 In total we found 126 singular readings, including seven places where the original text was illegible because of a rewriting. These are all included in the category “Nonsense word forms.” Analytic Category

Orthography Nonsense word forms Leaps a/o verba minora

Scribe D number/10 folios (real number)

5.2 (1)

Scribe A number/10 folios (real number) 5.5 (8)86 12.4 (18) 4.1 (6) 10.4 (15)

84 The number of folios has been adjusted to take account of the final columns of a book, which were only partially used. 85 One reading was excluded because supporting patristic evidence emerged whilst researching the particular reading: according to the TLG, Epiphanius, Basilius Seleuciensis, and Oecumenius all cite Col 2:14 with the reading upenantion hmwn (with Sinaiticus) instead of the normal upenantion hmin. 86 Excluding the reading kencraiais for kegcreais because the only issue here is assimilation of -n- to -g-.

203

Scribes D and A in Paul Analytic Category

Harmonisation - immediate context - intermediate context - general usage Editorial readings - interfering meaning - misreading syntax - misunderstood corrections - deliberate improvements Nonsense meanings Substitutions Transpositions a/o words or clauses More complex rewritings Total:

Scribe D number/10 folios (real number)

Scribe A number/10 folios (real number) 9.0 (13) 4.8 (7) 1.4 (2) 2.8 (4) 2.1 (3)

15.6 (3) 15.6 (3) 36.5 (7)

0.7 (1) 1.4 (2) 7.6 (11) 12.4 (18) 4.8 (7) 11.7 (17) 2.1 (3) 82.1 (119)

Of the seven singular readings by scribe D, two have been corrected; in scribe A’s section 70 of the 119 singular readings were corrected by the earliest correctors. Unlike the situation in 1 Chronicles, where it seems certain that all the corrections are made against the exemplar, a case can be made for the involvement of a second manuscript in the correction phase in the Pauline letters. The existence of such a second manuscript is likely when corrections of non-singular readings having support from a manuscript or group of manuscripts are made into the direction of another manuscript or group of manuscripts. The following six examples from Romans may serve to show that the original reading of Sinaiticus has support from cognate manuscripts (A 1739) but that likewise the corrections, made by the earliest correctors, have good support.87 1. Rom 1:28: omit o qeos with A 0172*, corrected towards the rest of the tradition. 2. Rom 4:2: propatora with A B C* 6 81 365 1506 (1739), corrected to patera with C3 D F G Y 629 (and with a different word order of the whole clause 33 1881 Maj). 3. Rom 5:1: ecwmen with A B* C D K L 33 81 630 1175 1739* pm, corrected to ecomen with B2 F G P Y 0220vid pm. 87

The witnesses to each reading are taken from NA27.

204

Singular Readings and Individual Scribal Patterns

4. Rom 5:7: mogis with 1739 pc, corrected to molis with the rest of the tradition. 5. Rom 7:25: eucaristw tw qew with A 1739 1881 Maj, corrected to caris de tw qew with B (which omits de) Y 33 81 104 365 1506. 6. Rom 9:27: upoleimma with A B 81 1739c, corrected to kataleimma with P46 D F G Y 33 1739* 1881 Maj. 7. Rom 10:17: cristou with P46(vid) B C D* 6 81 629 1506 1739, corrected to qeou with A D1 Y 33 1881 Maj. Though no clear pattern arises from these examples, they suffice to suggest that the correction manuscript stood reasonably close to the (secondary) Alexandrian tradition. Therefore, the link between correction and scribe-created reading which existed in 1 Chronicles and also in Psalms, where all the corrections were probably made impromptu, cannot be assumed for the Pauline corpus. Discussion of the singular readings per category

Orthography Rom 2:25, folio 81.1 (NT 62), line 7.43 Rom 5:3, folio 81.2 (NT 63), line 7.4 Rom 7:15, folio 81.3 (NT 64), line 7.11 Rom 9:12, folio 81.4 (NT 65), line 6.24 Rom 11:1, folio 81.5 (NT 66), line 3.14 Col 1:7, folio 83.8 (NT 85), line 8.38 2 Thess 3:8, folio 84.4 (NT 89), line 5.44 1 Thess 1:8, folio 84.2 (NT 87), line 6.7

ℵ akrob[inc]i[/inc]stia

exemplar akrobustia

caucwmeqa

kaucwmeqa

alla

all

meizon

meizwn

isdrahleiths

israhliths

emaqate

emaqete

alla

all

alla n

alla en

Twice we find alla instead of the apocopated form all j, and once the form allan for alla en at the end of a line (the -n- is written as a bar above a: alla( ). This is another instance of inverse elision, which is relatively uncommon in Sinaiticus but very common in

205

Scribes D and A in Paul

popular language.88 The i for u in akrobustia was immediately corrected back, even before the scribe wrote -stia. The spelling isdrahleiths is remarkable and occurs elsewhere in Sinaiticus, to my knowledge, only in Acts.89 The form emaqate for emaqete is an instance of the common intrusion of weak aorist forms into the strong aorist.90

Nonsense word forms Rom 1:12, folio 81.1 (NT 62), line 2.8 Rom 1:16, folio 81.1 (NT 62), line 2.34 Rom 3:31, folio 81.2 (NT 63), line 3.30 Rom 5:16, folio 81.2 (NT 63), line 8.46 Rom 8:7, folio 81.4 (NT 65), line 1.12 Rom 10:1, folio 81.4 (NT 65), line 8.43 Rom 15:12, folio 81.6 (NT 67), line 7.45 Col 2:3, folio 84.1 (NT 86), line 4.30 Col 2:8, folio 84.1 (NT 86), line 5.7 Col 2:14, folio 84.1 (NT 86), line 6.1 Col 2:23, folio 84.1 (NT 86), line 7.8 Col 4:8, folio 84.2 (NT 87), line 2.43 Heb 1:1, folio 84.4 (NT 89), line 8.4

ℵ en n

exemplar en

ioude

ioudaiw te

[inc]n[/inc]h

mh

amarthtos

amarthsantos

ounatai

dunatai

[ill]erasure[/ill]

pros ton qn+

anistanomenos

anistamenos

[ill]erasure[/ill]sews

gnwsews

sulagwn

sulagwgwn

[ill]1 letter[/ill]ois

tois

[ill]erasure[/ill] ts

crhstothta tapeinofrosunh tas

pat[ill]erasure[/ill]

patrasin

For inverse elision, or aphaeresis, see Gignac I, 319–21. In Latin manuscripts the spelling of Israel with a t (istrael) or d (isdrael) is more common. In Greek it is very rare, though it does occur in Vaticanus. See Milne-Skeat, 66–67. 90 Moulton-Howard, 208; Gignac II, 335. 88 89

206

Singular Readings and Individual Scribal Patterns

Heb 2:7, folio 84.5 (NT 90), line 2.40 Heb 10:34, folio 84.8 (NT 93), line 3.14 Heb 11:3, folio 84.8 (NT 93), line 4.5 Heb 12:4, folio 85.1 (NT 94), line 2.6 Heb 13:18, folio 85.1 (NT 94), line 7.24

ℵ estefanwsa[ill]erasu re[/ill] ginwskon

exemplar estefanwsas auton ginwskontes

fain

fainomenwn

ant[ill]3 letters[/ill]izomenoi oti kalhnqa

antagwnizo-menoi peiqomeqa

The 18 nonsense forms include the seven instances of rewritten text in which the original text is illegible. In six cases the reading in this category is shorter than it should be because either a single letter (once) or one or more syllables have fallen away. In only one instance is a syllable added, anistanomenos for anistamenos in Rom 15:12. The added syllable is the first of a new line. In Rom 1:12 the initial letter e of the preposition en occupies the last position of the line. Above it we find the crossbar, indicating a final nu, but also on the next line the same letter is written plene.91 The influence of the immediate context can result in straight nonsense words as the following example shows: Heb 13:18, folio 85.1 (NT 94), lines 7.23–27 proseucesqe pe ri hmwn oti kalh( qa gar oti kalhn sunidhsin eco men

Instead of writing peiqome - qa the scribe produces otikalh( qa, which is actually what he has to write at the end of the third line. The nu being written as a crossbar, the resultant text has the same number of letters as the intended text. Surprisingly, this clear error has not been corrected in the scriptorium.

Leaps from the same to the same ℵ Rom 4:12, folio 81.2 (NT peritomhs 63), line 5.2 91

exemplar peritomhs ois ouk ek peritomhs

It is not uncommon to break off a proclitic as e–n.

207

Scribes D and A in Paul

Rom 14:23, folio 81.6 (NT 67), line 6.28 Col 2:18, folio 84.1 (NT 86), line 6.24 1 Thess 2:14, folio 84.2 (NT 87), line 8.34 Heb 1:8, folio 84.5 (NT 90), line 1.7 Heb 9:12, folio 84.7 (NT 92), line 4.33

twn mellontwn

exemplar pistews pan de o ouk ek pistews twn

see text

see text

h rabdos

h rabdos ths euquthtos rabdos efapax eis ta agia

ℵ pistews

eis ta agia efapax eis ta agia

The six leaps in the Pauline section can be divided into leaps forward and backward. Interestingly, all of these have been corrected, either by the scribe as prima manus or by scribe A or D acting as corrector. In Rom 4:12 the scribe jumps from peritomhs to peritomhs missing out tois ouk en; Rom 14:23 jumps from ouk ek pistews to the same words later in the verse, skipping pan de o; and Heb 1:8 omits the words rabdos ths euquthtos, a leap from rabdos to rabdos. To these three cases of a leap forward, three leaps backward can be added. Colossians 2:18 has the unique reading qrhskeia twn mellontwn aggelwn. The intrusion of mellontwn comes from the previous verse, where the text reads skia twn mellontwn. Apparently we have to do here with an aborted leap back triggered by sk(e)ia twn. In Heb 9:12 we have another leap back which may have been caused by an initially forgotten efapax: eis ta agia efapax eis ta agia. The first clause is marked for deletion by the first hand.92 The most substantial leap backward is found at 1 Thess 2:14, just before the replacement leaf by scribe D: 1 Thess 2:14, folio 84.2 (NT 87), lines 8.24–44 First time

Second time

24 akohs par hmwn 25 tou qu+ edexasqe ou 26 logon anqrwpwn

34 wn tou qu+ edexa35 sqe ou logon an36 qrwpwn alla ka-

92 In cases such as the variant in Heb 9:12, one can equally well label the reading as an in-line correction of an omission, a word order variant, or an aborted leap back.

208

Singular Readings and Individual Scribal Patterns

First time 27 alla kaqws esti( 28 logon qu+ os kai e29 nergeitai en hmi( 30 tois pisteuousin 31umis gar mimhtai 32 egenhqhte adel33 foi twn ekklhsi34 wn tou qu+ edexa-

Second time 37 qws estin logon 38 qu+ os kai energi39 tai en hmin tois 40 pisteuousin 41umeis gar mimhqh42 te egenhqhte adel43 foi twn ekklhsi44 wn tou qu+ twn ou-

The leap backward is triggered by the letters -wn tou qu+. Interestingly, we have a number of differences between the two versions: energeitai versus energitai (lines 28–29 and 38–39), umis versus umeis (lines 31 and 41), and mimhtai versus mimhqhte (lines 31 and 41–42). The differences in numbers 1 and 2 are simple orthographic interchanges. In number 3 we have a replacement of mimhtai with mimhqhte. Though it is possible to explain this form as a combination of an interchange of ai–e with an extra syllable -qh-, it is perhaps more easily explained by harmonisation to the next word with the same ending, egenhqhte. The form mimhqhte is a true singular reading and, given the specific nature of this large leap backwards, very probably a scribal creation. Besides these three differences, both versions of the repeated passage contain two readings only found in Sinaiticus and are therefore labelled as singular readings. The reading kaqws estin (lines 27 and 36/37) is a singular reading for the normal kaqws alhqws estin. alhqws may have dropped out because of homoeoteleuton. The second reading is en hmin (lines 29 and 39) for the normal en umin.

Add and omit verba minora exemplar ℵ 1 Thess 5:1, folio 84.3 tou grafesqai umin grafesqai umin (NT 88), line 5.43 Rom 4:13, folio 81.2 tw spermati (NT 63), line 5.10 Rom 10:16, folio 81.5 en tw euaggeliw (NT 66), line 2.24

h tw spermati tw euaggeliw

Scribes D and A in Paul

Rom 11:31, folio 81.5 (NT 66), line 6.38 Rom 16:7, folio 81.7 (NT 68), line 3.31 Col 2:18, folio 84.1 (NT 86), line 6.23 Col 3:1, folio 84.1 (NT 86), line 7.16 2 Thess 1:3, folio 84.4 (NT 89), line 1.19 2 Thess 2:16, folio 84.4 (NT 89), line 4.41 2 Thess 3:10, folio 84.4 (NT 89), line 6.8 Heb 1:5, folio 84.4 (NT 89), line 8.36 Heb 1:12, folio 84.5 (NT 90), line 1.31 Heb 4:11, folio 84.5 (NT 90), line 8.14 Heb 11:9, folio 84.8 (NT 93), line 5.16 Heb 12:10, folio 85.1 (NT 94), line 2.43 Heb 13:22, folio 85.1 (NT 94), line 8.9

ℵ nun elehqwsin kai tapinofrosunh

209

exemplar autoi nun elehqwsin oi kai

omitted

en tapeinofrosunh estin

pantwn

pantwn umwn

o qeos

kai o qeos

parhggellomen esomai

touto parhggellomen esomai autw

de kai

de

pesh

tis ... pesh

ths epaggelias auths metalabein

ths epaggelias ths auths eis to metalabein

kai

kai gar

In total we find 16 instances of the addition or omission of verba minora. Four times an addition took place: the preposition en was added to the dative tw euaggeliw in Rom 10:16 (corrected by the prima manus), and a genitive article was added to the infinitive in 1 Thess 5:1, which is the only instance of this category of singular readings in scribe D’s text. Noteworthy is the change in meaning that arose because of the omission of the second ths in Heb 11:9: ths epaggelias ths auths. “The same promise” has become “her promise”. The correct meaning was restored by changing the word order ths epaggelias auths to ths auths epaggelias, though the specific form of the correction is a new singular reading.93 Finally, in Heb 1:12, the scribe Tischendorf mentions also other readings: tou epaggelia autou; ta epaggelia autwn, and ta epaggelia. 93

210

Singular Readings and Individual Scribal Patterns

added kai to de only to correct it himself. Of the twelve omissions, seven have been corrected.

Harmonisations Harmonisation to the immediate context ℵ Rom 9:32, folio prosekoyen 81.4 (NT 65), line 8.29

exemplar prosekoyan

Rom 11:22, folio crhstoth-tos crhstoths 81.5 (NT 66), line qeou qeou 5.36

Rom 15:28, folio sfragisame81.7 (NT 68), line nois autois 2.6 ina Col 4:3, folio 84.2 (NT 87), line 2.15 pirasmw Heb 3:8, folio 84.5 (NT 90), line 5.22 Heb 10:26, folio ths 84.8 (NT 93), line epignwsian 2.10

sfragisamenos autois ama

parapikrasmw thn epignwsin

Israel kept as referent instead of switching to the plural Grammatical anomaly under influence of two other occurrences of inflected crhstoths in same verse (with the extra syllable) Following word is autois May have been corrected straightaway. Upcoming ina 4 words later Influence of upcoming word epignwsian is a non-existent form but created under the influence of the following ths alhqeias

Scribes D and A in Paul ℵ Heb 12:10, folio o men 85.1 (NT 94), line 2.39

exemplar oi men

211

Verb epaideuon is plural. Preceding “father of spirits” is singular. Either “father of spirits” is taken as referent, or harmonisation to o de

Harmonisation to the intermediate context ℵ Rom 14:20, folio apollue 81.6 (NT 67), line 6.9 Heb 10:32, folio amartias 84.8 (NT 93), line umwn 2.45

exemplar katalue

Verb used in 14:15

hmeras

Harmonisation to remote context / other passages ℵ Rom 15:9, folio kaqws 81.6 (NT 67), line gegraptai dia 7.29 tou profhtou Rom 16:25, folio to kuriou 81.7 (NT 68), line ihsou cristou 5.38 Col 3:1, folio sunhgerqh-te 84.1 (NT 86), line en cristw 7.15 Heb 12:20, folio quk 85.1 (NT 94), line 4.6

exemplar kaqws gegraptai dia touto to khrugma ihsou cristou sunhgerqh-te en cristw tw cristw common expression; cf. also Rom 6:11; 1 Thess 4:16 ouk Initially read as logon qu instead of logon ouk. No crossbar: correction was made while writing the text.

212

Singular Readings and Individual Scribal Patterns

Seven times a singular reading arose because of harmonisation to the immediate context. Three of these involve the singular or plural declension of the word: prosekoyan in Rom 9:32 was made singular in continuation of the previous verse; in Rom 15:28 sfragisamenos was harmonised to the following word autois; and in Heb 12:10 the scribe commenced the first half of the verse with o men instead of oi men, just as the second half opens with o de. (The last case was corrected by an early corrector.) The form ths epignwsian in Heb 10:26 was written under the strong influence of the following ths alhqeias, the normal reading being thn epignwsin; both the case of the article and the shape of the noun suffix underwent influence from the following word.94 In Rom 11:22 the non-grammatical form crhstothtos is found (for crhstoths). This is probably not simply a nonsense form but arose under the influence of crhstothta and crhstothti elsewhere in the same verse. Twice a word was exchanged for another word taken from the immediate context. In Col 4:3 ama is replaced with ina, which occurs four words further down, and in Heb 3:8 the text reads ws en tw pirasmw kata thn hmeran tou pirasmou, where tw pirasmw replaces tw parapikrasmw. Four of the harmonisations to the immediate context arose because the text was influenced by words that followed the actual singular reading. Influence from the intermediate context can be seen in Rom 14:20, where the reading apollue for katalue may have been influenced by the occurrence of alollue in Rom 14:15. In Heb 10:32, Sinaiticus has the singular reading anamimnhskesqe de tas proteron amartias umwn instead of the normal proteron hmeras. It may well be that this is due to the influence from the quotation of Jer 31:34 in Heb 8:12 and 10:17, where it is said that twn amartiwn autwn ou mh mnhsqhsomai/mnhsqw. The main verb in these two passages is almost identical to Heb 10:32, which may have been the trigger for the substitution of hmeras with amartiwn umwn. Above, on page 154, we have seen a similar influence of the Jeremiah text on scribe A and on how he phrased the promise to David in 1 Chr 17.

epignwsian was corrected by an A corrector (and later also by a C corrector) to epignwsin, ths was only corrected by a late C corrector. There is no comment by Milne and Skeat in the notebook. 94

213

Scribes D and A in Paul

On four occasions singular readings are found that are best explained by the influence of general biblical style and usage. The quotation in Rom 15:9 is introduced with the standard formula kaqws gegraptai after which the first words of the quotation read dia touto. In Sinaiticus the uncorrected text reads kaqws gegraptai dia tou profhtou. In Rom 16:25 we find the expression to euaggelion mou kai to khrugma ihsou cristou. The scribe made this into to euaggelion mou kai ku+ iu+ cu+. The expression sunhgerqhte tw cw+ becomes sunhgerqhte en cw+ in Col 3:1, possibly under the influence of passages like 1 Thess 4:16. The last example of harmonisation to general usage is found in Heb 12:20. The last word of the previous verse is logon, after which the scribe continued with qu, which, as a nomen sacrum would be a natural continuation. Before the crossbar was placed above these letters, the error was noted and restored to ouk. All the singular readings made under the influence of general usage have been corrected by the prima manus or a corrector.

Editorial readings 1 Thess 1:5, folio 84.2 (NT 87), line 5.28 Heb 10:12, folio 84.7 (NT 92), line 8.35 Rom 16:2, folio 81.7 (NT 68), line 3.2

ℵ euaggelion tou qeou hmwn ek dexia

exemplar euaggelion tou qeou en dexia

kai autou kai emou kai autou emou

The reading of Sinaiticus in 1 Thess 1:5 may well be the result of a misunderstood correction in the exemplar. Sinaiticus reads to euaggelion tou qu+ hmwn while the overwhelming majority of the tradition has to euaggelion hmwn. However, Ephraemi Syri Rescriptus reads to euaggelion tou qu+, which opens up the possibility that Sinaiticus shows a conflation of these two readings. Similarly, a conflation of two readings may have taken place in Heb 10:12. As it stands, the reading ek dexia does not make sense, but if it is understood as a conflation of en dexia, as read in the majority of the tradition, and ek dexiwn (in itself a harmonisation towards Ps 109 [LXX]:1), as found in Alexandrinus and minuscule 104, the reading becomes understandable. The text of Rom 16:2 shows considerable variation around the words kai emou autou. Sinaiticus reads kai emou kai autou, which

214

Singular Readings and Individual Scribal Patterns

barely makes sense in the context, but resolves the tension that might arise if autou in the original wording is not recognised as intensive.95

Nonsense meanings in context Rom 4:18, folio 81.2 (NT 63), line 5.48 Rom 5:9, folio 81.2 (NT 63), line 7.36 Rom 6:22, folio 81.3 (NT 64), line 4.43 Rom 11:8, folio 81.5 (NT 66), line 4.4 2 Thess 2:7, folio 84.4 (NT 89), line 3.30 Heb 2:4, folio 84.5 (NT 90), line 2.24 Heb 4:14, folio 84.5 (NT 90), line 8.36 Heb 8:8, folio 84.7 (NT 92), line 1.48 Heb 10:36, folio 84.8 (NT 93), line 3.25 Heb 10:39, folio 84.8 (NT 93), line 3.38 Heb 13:12, folio 85.1 (NT 94), line 6.32

ℵ kata kata

exemplar kata

en nun en

nun en

te

de

oo

o

to gar musterion to gar musterion hdh gar energeitai hdh energeitai qerismois merismois dielhluqa

dielhluqota

oi

oikon

komisasqai

komishsqe

apwlias

apwleian

omitted

epaqen

In total, 11 singular readings are classified under the category of nonsense meanings in context. Most of these are substitutions or dittographic errors. Others are formed by omitting a syllable. Six of these have been corrected by the scribe or a corrector.

Substitutions 1 Thess 3:11, folio 84.3 (NT 88), line 3.7 1 Thess 4:10, folio 84.3 (NT 88), line 4.27

ℵ umwn

exemplar hmwn

adelfwn umwn en olh th makedonia

adelfwn tous en olh th makedonia

This singular reading might equally have been classified under “nonsense meanings.” 95

215

Scribes D and A in Paul

1 Thess 4:14, folio 84.3 (NT 88), line 5.4 Rom 5:5, folio 81.2 (NT 63), line 7.15 Col 1:19, folio 84.1 (NT 86), line 2.21 Col 1:27, folio 84.1 (NT 86), line 3.39 Col 2:18, folio 84.1 (NT 86), line 6.30 Col 3:1, folio 84.1 (NT 86), line 7.16 Col 3:10, folio 84.1 (NT 86), line 8.8 Col 4:12, folio 84.2 (NT 87), line 3.22 1 Thess 2:9, folio 84.2 (NT 87), line 7.47 1 Thess 2:13, folio 84.2 (NT 87), line 8.29 2 Thess 1:7, folio 84.4 (NT 89), line 1.43 2 Thess 2:14, folio 84.4 (NT 89), line 4.30 Heb 2:8, folio 84.5 (NT 90), line 2.45 Heb 3:6, folio 84.5 (NT 90), line 5.11 Heb 8:10, folio 84.7 (NT 92), line 2.26 Heb 10:18, folio 84.8 (NT 93), line 1.14 Heb 11:5, folio 84.8 (NT 93), line 4.21 Heb 13:2, folio 85.1 (NT 94), line 5.17 Heb 13:23, folio 85.1 (NT 94), line 8.15

ℵ episteuomen

exemplar pisteuomen

umwn

hmwn

edokhsen

eudokhsen

tou

toutou

sarkos autwn

sarkos autou

qeos

cristos

ependusamenoi

endusamenoi

hmwn

umwn

umin

eis umas

hmin

umin

umwn

hmwn

umwn

hmwn

upetaxa

upetaxas

kan

ean

mou

moi

afeis

afesis

oti

dioti

thn filoxenian

ths filoxenias

erchsqe

erchtai

Scribe D has three substitutions, all on his first replacement folio. Once umwn is replaced with hmwn, once tous with hmwn in 1

216

Singular Readings and Individual Scribal Patterns

Thess 4:10,96 and finally the present pisteuomen was made into an imperfect.97 Among the 18 substitutions by scribe A, we find on five occasions a change from umeis to hmeis or vice versa (three and two times, respectively). Four times we find changes in the verb form, some of them only minor: edokhsen for eudokhsen, ependusamenoi for endusamenoi, upetaxa for upetaxas and erchsqe for erchtai. Only twice has the original word been replaced with a different word: cs+ with qs+, and ean with kan, but even in these two instances the change is minimal.

Transpositions Rom 9:3, folio 81.4 (NT 65), line 5.20 Rom 10:5, folio 81.5 (NT 66), line 1.14 Rom 15:32, folio 81.7 (NT 68), line 2.29 Col 2:20, folio 84.1 (NT 86), line 6.40 Heb 4:7, folio 84.5 (NT 90), line 7.37 Heb 10:11, folio 84.7 (NT 92), line 8.22 Heb 10:36, folio 84.8 (NT 93), line 3.23

ℵ einai anaqema autos egw oti thn dikaiosunhn thn ek nomou elqwn en cara

exemplar anaqema einai autos egw thn dikaiosunhn thn ek nomou oti

apoqanete oun

oun apeqanete

orizei tina

tina orizei

litourgwn kaq hmeran crian ecete

kaq hmeran leitourgwn ecete creian

en cara elqwn

In this particular instance it is difficult to determine the text in the exemplar. tous adelfous en olh is read by A D* F G 629 lat, while tous adelfous tous en olh is read by most other witnesses. Both readings could underlie the reading of Sinaiticus: tous adelfous hmwn en olh. On the weight of the external evidence the majority reading is preferred. If not, this singular reading ought to be reclassified as an addition of a small word. 97 Tischendorf assigns the correction of the imperfect to a C corrector, and initially I had some doubt. On inspection it turned out that the correction is indeed similar to other C corrections such as on the next page folio 84.4 (NT 89), line 4.41. 96

Scribes D and A in Paul

217

The seven transpositions by scribe A are mostly between two words or two clauses. The only exception is Rom 10:5, where Sinaiticus has a transposition of oti over five words and two clauses.

Add and omit words and clauses 1 Thess 5:8, folio 84.3 (NT 88), line 6.34 Heb 7:10, folio 84.6 (NT 91), line 6.25 Heb 7:21, folio 84.6 (NT 91), line 7.35 Col 4:2, folio 84.2 (NT 87), line 2.13 Rom 1:8, folio 81.1 (NT 62), line 1.33 Rom 2:16, folio 81.1 (NT 62), line 7.2 Rom 4:12, folio 81.2 (NT 63), line 5.5 Rom 11:30, folio 81.5 (NT 66), line 6.35

Rom 15:2, folio 81.6 (NT 67), line 6.36 Col 2:1, folio 84.1 (NT 86), line 4.17 Col 2:7, folio 84.1 (NT 86), line 5.1 Col 4:7, folio 84.2 (NT 87), line 2.38 1 Thess 2:13, folio 84.2 (NT 87), line 8.28 2 Thess 2:4, folio 84.4 (NT 89), line 3.14 2 Thess 2:16, folio 84.4 (NT 89), line 4.42 Heb 2:18, folio 84.5 (NT 90), line 4.25

ℵ omitted

exemplar kai agaphs

omitted

eti gar

omitted

eis ton aiwna

omitted

en auth

omitted

dia ihsou cristou

cristou ihsou

dia cristou ihsou

ths en akrobustia omitted

omitted

ths en akrobustia pistews wsper gar kai umeis pote hpeiqhsate tw qew nun de hlehqhte th toutwn apeiqeia eis to agaqon

omitted

en sarki

omitted

en autw

omitted

kai sundoulos

omitted

alhqws

omitted

kai uperairomenos

omitted

o agaphsas hmas

omitted

peirasqeis

218

Singular Readings and Individual Scribal Patterns exemplar ara apoleipetai sabbatismos tw law tou qeou acri merismou kai acri merismou pneumatos yuchs kai pneumatos hkw omitted

ℵ Heb 4:9, folio 84.5 (NT omitted 90), line 8.2 Heb 4:12, folio 84.5 (NT 90), line 8.22 Heb 10:7, folio 84.7 (NT 92), line 7.44 Heb 12:21, folio 85.1 (NT 94), line 4.14

omitted

eimi

In scribe D’s section we find three singular omissions: kai agaphs in the expression qwraka pistews kai agaphs, eti gar at the start of Heb 7:10, and the clause eis ton aiwna in Heb 7:21. The last two omissions were repaired. Scribe A shows 17 variants in this category, all of them omissions as well. Most of these omissions do not disturb the flow of the sentence except the omission of pistews in Rom 4:1298 and the omission of yuchs in Heb 4:12. The other 15 variants show omissions of between one and seven words.99

Major rewritings Col 2:10, folio 84.1 (NT 86), line 5.23 Heb 11:8, folio 84.8 (NT 93), line 5.5 Heb 12:23, folio 85.1 (NT 94), line 4.26

ℵ pashs ths archs ekklhsias klhronomian lambanin teliwn dedikaiwmenois

exemplar pashs archs kai exousias lambanein eis klhronomian dikaiwn teteleiwme-nwn

In the three major rewritings, a combination of phenomena happens. In Heb 11:8 the preposition before the noun is dropped and the noun and infinitive swap places, klhronomian lambanin for lambanein eis klhronomian. This reading is corrected towards the normal text. In Heb 12:23, the verb becomes noun and the noun verb: dikaiwn teteleiwmenwn changes into teliwn 98 Note that earlier in this verse a leap from peritomhs to peritomhs occurs (see above, page 207). 99 Seven times one word, four times two words, three times three words, once four words, once seven words.

Scribes D and A in Paul

219

dedikaiwmenois. The correctors did not touch this singular reading. The third reading is found in Col 2:10. The normal text reads h kefalh pashs archs kai exousias and is changed to h kefalh pashs ths archs ekklhsias. The substitution of kai exousias for ekklhsias was most probably triggered by the association of kefalh and ekklhsia in Col 1:18 and Eph 1:22. The noun clause pashs archs now became qualified by ekklhsias and received an article, “head of every ruler (or “of each office”) of the church.” Conclusions The difference in the number of singular readings between the folios written by scribe D and by scribe A is very large. Scribe D has only a fraction of the number of singular readings that scribe A has. It is difficult to imagine that scribe D uses a different exemplar for the two replacement folios than that used by scribe A. Likewise, it is not probable that he uses scribe A’s version of the text as a basis for the replacement, as, in that case, one would have expected the inclusion of singular readings created by scribe A. We have therefore a copy of a small part of the exemplar by scribe D and of a large part of the same exemplar by scribe A. The substantial difference between the frequency of singular readings can therefore only be attributed to a difference in the personal scribal characteristics of the two scribes. Though the sample from scribe D is relatively small, the fact that we found a similar difference between the scribes in Psalms adds credibility to these results from Paul. The special nature of the replacement sheets shows up very well in the absence of any nonsense readings or meanings. The two folios have been written very carefully with no clear mistakes in the text. The two singular readings on folio 84.6 (NT 91), containing the text from Hebrews, have been corrected by the prima manus, but, with a possible exception in 1 Thess 4:14, no corrections have been added at all to folio 84.3 (NT 88). In our analysis of scribe A, we have so far assumed that the various parts of the Pauline corpus could be compared well with one another and therefore we have given the averages over the whole sample. Interestingly though, a difference can be seen in the number of singular readings per folio between Romans and the section 1 Thessalonians–Hebrews.

220

Singular Readings and Individual Scribal Patterns

20 15 10 5

85 .1

84 .7

84 .5

84 .3

84 .1

81 .7

81 .5

81 .3

81 .1

0

Romans is written on folios 81.1 (NT 62)–81.7 (NT 68); 1 Thessalonians starts on the verso of folio 83.8 (NT 85; this half folio is left out from the graph above).100 The difference between the figures for Romans and 1 Thessalonians–Hebrews is not hard to notice; neither is it difficult to spot the two replacement folios by scribe D in the second block (folio 84.3 and 84.6). In the previous chapter, when discussing the number of paragraphs in the Pauline corpus,101 we noticed that Romans contained more paragraphs than the other letters in the Pauline corpus. Could there be a relation between a high number of paragraphs per folio and a low number of singular readings? Such relation may be understandable when both phenomena are seen as functions of the same basic attitude, namely that of taking good care of the text. At places where the scribe makes an effort to start each thought or sentence on a new line, the scribe will make fewer casual errors because of the increased level of concentration.102 The details of the large leap backwards in 1 Thess 2 have been discussed above (page 207). However, we did not address the possible implications of the singular readings in this leap for the study of singular readings as a whole. Two of the singular readings in the first version are repeated in the second version. What is the likelihood of the scribe making the same mistake twice when he erroneously repeats a passage? Is this not an argument to consider the majority of the singular readings as being copied errors rather than scribe-created errors? Though the argument has a certain force, the circumstances of Sinaiticus point in a different direction. 100 The first four columns of 1 Thessalonians contain only one singular reading which is, certainly for scribe A’s standards a very low number. 101 See the discussion in Chapter 3, page 98. 102 Alternatively, a paragraph break can be seen as a pause in the activity of the scribe, which may also help concentration.

Scribes D and A in Luke

221

First, one can never be sure that any particular singular reading is indeed created by the scribe, simply because we do not possess the exemplar from which the scribe worked. Only in the case of a pattern of singular readings we can be more confident. Secondly, given the fact that different scribes can be detected when one looks at the number and pattern of singular readings (at least when it comes to Sinaiticus), it becomes clear that though perhaps not each singular reading is scribe-created, the majority of these readings certainly are. Thirdly, it is by no means improbable that a scribe makes the same mistake twice within a short interval. Once one has read a particular sentence in a certain way it is not strange to assume that such a reading will linger on in the memory of a scribe and that therefore the same mistake is quite likely to be repeated.

SCRIBES D AND A IN LUKE Introduction Our final study of singular readings concentrates on the first 12 chapters of Luke.103 This section contains a replacement folio by scribe D, with the majority of the text written by scribe A. This enables us once again to compare the work of two scribes working on the same book. In addition, this is the only Gospel for which a recent and reliable apparatus is available in the volumes of the IGNTP. The folio of scribe D belongs to the replacement sheet covering the last folio of Mark and the first of Luke. The total text written by scribe D is 0.75 folio, while scribe A wrote the remaining 9.87 folios.104 As with our study of a selection of the Pauline corpus, it is relevant to ascertain the status of the corrections and whether indications exist to assume the use of a second manuscript in the correction phase. The following readings by the earliest correctors taken from the first four chapters suffice: 1. Luke 3:15 autwn with the whole Greek tradition, corrected to eautwn, a singular reading. 2. Luke 3:17 diakaqare ... sunagagein with P4 B, corrected to kai diakaqariei ... sunaxei with the rest of the tradition.

103 The first 12 chapters of Luke comprise Volume 1 of the IGNTP, The Gospel according to St. Luke, (Oxford: Clarendon, 1984). 104 Luke 12 ends on folio 77.7 (NT 39), line 7.36.

222

Singular Readings and Individual Scribal Patterns

3. Luke 4:5 kai anagagwn auton with B L 1241 pc, corrected to kai anagagwn auton eis oros uyelon with (D W) f1 700 2542 l 844 l 2211 pc and also with many others who have o diabolos before eis oros uyelos. 4. Luke 4:36 exercontai with the whole Greek tradition, corrected to upakouousin autw, a singular reading. 5. Luke 4:43 euaggelion tou qeou with 892, corrected to basileian tou qeou with rest of the tradition. Twice we see the creation of a singular reading, in Luke 3:15 and 4:36. The first correction is assigned by Tischendorf to corrector A. It may well be that this is the same hand who, in the earliest chapters of Matthew, was responsible for the creation of two similar singular readings, namely autou corrected to eautou in Matt 1:24 and 4:21. However, in both of those cases Tischendorf attributes the correction to corrector B.105 The corrections in Luke 3:17 and 4:5 alter a reading shared by a small group of manuscripts with a related text to a reading that is much more widespread. The correction of Luke 4:43 changes a reading originally shared with minuscule 892 only. Elsewhere Sinaiticus shares more readings with this manuscript making it less likely that this particular agreement is coincidental and the result of independently creating the same error.106 Because of these readings it seems hazardous to see the corrections in Luke as corrections of scribe-created readings. Apparently, more complex processes are at work. For the numbers in the following table a similar caveat is true as for the sample from Paul: scribe D is only represented by six columns of text, less than a folio. Inevitably, in converting the figures to numbers per 10 folios, a considerable distortion takes place. However, for the sake of comparison with the other studies in this chapter, we still give the figures in the same format as used in those studies. The numbers of singular readings for the first 12 chapters of Luke are as follows: Analytic Category

Orthography Nonsense word forms

Scribe D number/10 folios (real number) 26.7 (2)

Scribe A number/10 folios (real number) 11.1 (11) 20.3 (20)

105 Milne and Skeat (p. 45) have argued for the identity of corrector B with scribe A. 106 See J. Rendel Harris, “An Important Ms. of the New Testament (cod. Mus. Brit. Add. 33277)” Journal of Biblical Literature 9 (1890): 31–59.

223

Scribes D and A in Luke Analytic Category

Scribe D number/10 folios (real number)

Leaps a/o verba minora Harmonisation - immediate context - intermediate context - general usage Editorial readings - interfering meaning - misreading syntax - misunderstood corrections - deliberate improvements Nonsense meanings Substitutions Transpositions a/o words or clauses More complex rewritings Total:

Scribe A number/10 folios (real number) 1.0 (1) 19.3 (19) 15.2 (15) 9.1 (9) 3.0 (3) 3.0 (3) 2.0 (2)

26.7 (2) 26.7 (2)

2.0 (2) 9.1 (9) 19.3 (19) 5.1 (5) 21.3 (21) 2.0 (2) 125.6 (124)

26.7 (2)

80.0 (6)

Discussion of the singular readings per category

Orthography Luke 1:1, folio 76.5 (NT 29), line 3.2 Luke 1:4, folio 76.5 (NT 29), line 3.16 Luke 2:46, folio 76.7 (NT 31), line 1.9 Luke 3:29, folio 76.7 (NT 31), line 5.47 Luke 3:32, folio 76.7 (NT 31), line 6.13 Luke 4:4, folio 76.7 (NT 31), line 7.15 Luke 5:3, folio 76.8 (NT 32), line 5.1 Luke 6:2, folio 77.1 (NT 33), line 2.31

ℵ epeceilhsan

exemplar epeceirhsan

epignois

epignws

ephrwtwnta

eperwtwnta

maqqaaq

maqqat (?)

balls

booz (?) bals

apokriqh

apekriqh

erwthsen

hrwthsen

ouc

ouk

224

Singular Readings and Individual Scribal Patterns

Luke 6:17, folio 77.1 (NT 33), line 4.36 Luke 9:3, folio 77.4 (NT 36), line 1.38 Luke 9:9, folio 77.4 (NT 36), line 2.26 Luke 9:62, folio 77.5 (NT 37), line 2.13 Luke 11:51, folio 77.6 (NT 38), line 7.23

ℵ ierousalhm kai piraias mhde

exemplar ierousalhm kai peraios (?) mhte (4)

akoiw

akouw

epibal[inc]o[/inc]n107

epibalwn

metoxu

metaxu

The two orthographic irregularities in scribe D’s text are epignois for the normal epignws and epeceilhsan instead of epeceirhsan. In the first case the iota subscript, which is not written in majuscule scripts, was made clear as an adscript with the omega reduced to an omicron. Though the iota adscript is common in papyri and inscriptions, it is non-existent or otherwise extremely rare in Sinaiticus.108 In epeceilhsan, which was already corrected by the prima manus, we have an example of the interchange between r and l which is often found in the papyri.109 In Sinaiticus we have found hardly any comparable instances of this interchange. In scribe A’s text, we twice find the omission of the augment before the verb in favour of the vowel that would take this position in the unaugmented form (Luke 4:4 and 5:3). The form metoxu for metaxu is frequently found in the papyri,110 and this is to a lesser extent also true for akoiw for akouw.111 In Luke 6:2, the form ouc in ouc exestin (instead of ouk) is difficult to explain. According to Gignac, there is little evidence for the interchange of aspirated and voiced stops,112 but on the other hand it is hard to imagine that the scribe interprets ex as a numeral.113 The tags [inc] and [/inc] enclose a letter that is uncertain. I have not found any examples of the iota adscript. Gignac I, 183 observes that the diphthong wi is usually written without the -i in the first and second centuries AD but was common thereafter. 109 Gignac I, 104. 110 Gignac I, 287. 111 Gignac I, 215. 112 Gignac I, 96. 113 It is possible, though, to argue that the numeral ex was in the scribe’s mind. The context is that some Pharisees say to Jesus ti poieite o ouk 107 108

Scribes D and A in Luke

225

Included amongst these orthographic readings is also the reading mhde for mhte (4) in Luke 9:3. The scribe commences a new line for each of the five items in the list and uses mhte for each one but the fourth. Strictly speaking mhte and mhde are two different lexical entries, but in all likelihood they should be treated in this context as mere spelling variants. Twice we find remarkable spellings of proper nouns in the genealogy of Luke 3. Instead of booz, Sinaiticus reads balls corrected by prima manus or an early corrector to bals, and for maqqat Sinaiticus has maqqaaq. It is impossible to establish what the scribe found in his exemplar as the whole tradition is littered with variants in the spelling of this name. A third remarkable spelling is the form piraias in Luke 6:17. The name of this region is spelled pereas in W, and peraias in some lectionaries and the Old Latin (it does not occur at all in the remainder of the Greek tradition). Gignac mentions the “occasional interchange” of e and i, which would explain the presence of the first iota in piraias.114 A possible example of an w–o interchange is found in Luke 9:62. Though the original version has been corrected by the prima manus or an early corrector and is barely legible, it is likely that epibalwn first read as epibalon. Another example will be seen below under the treatment of the verba minora, where the vocative particle w is written as o.

Nonsense words Luke 1:69, folio 76.6 (NT 30), line 2.19 Luke 2:12, folio 76.6 (NT 30), line 4.43 Luke 2:29, folio 76.6 (NT 30), line 7.6 Luke 2:42, folio 76.6 (NT 30), line 8.29

ℵ en n

exemplar en

essparganwmenon apollueis

esparganw-menon apolueis

otwn

etwn

exestin tois sabbasin in the context of a discussion about the Sabbath. In Luke 13:14 the ruler of the synagogue says in a similar context oti ex hmerai eisin en ais dei ergazesqai. Possibly this verse influenced the reading of Luke 6:2, which led the scribe to expect the numeral rather than the prefixed preposition. However, it remains simpler to explain ouc as a spelling anomaly, helped by the optical similarity of -c- and -k-. 114 Gignac I, 249; for this change before liquids see p. 250.

226

Singular Readings and Individual Scribal Patterns

Luke 3:1, folio 76.7 (NT 31), line 2.6

Luke 3:18, folio 76.7 (NT 31), line 4.31 Luke 5:13, folio 76.8 (NT 32), line 6.25 Luke 5:13, folio 76.8 (NT 32), line 6.27 Luke 7:6, folio 77.2 (NT 34), line 2.30 Luke 7:7, folio 77.2 (NT 34), line 2.35 Luke 8:15, folio 77.3 (NT 35), line 3.8 Luke 8:43, folio 77.3 (NT 35), line 7.43 Luke 9:21, folio 77.4 (NT 36), line 4.25 Luke 9:32, folio 77.4 (NT 36), line 6.1 Luke 9:39, folio 77.4 (NT 36), line 7.12 Luke 10:20, folio 77.5 (NT 37), line 4.40 Luke 11:36, folio 77.6 (NT 38), line 5.15 Luke 11:42, folio 77.6 (NT 38), line 6.2 Luke 11:42, folio 77.6 (NT 38), line 6.8 Luke 12:4, folio 77.6 (NT 38), line 8.28

euhggelize ton

exemplar itouraias kai (2) tracwnitidos cwras kai lusaniou euhggelizeto ton

egwn

legwn

euqews a

euqews

ikanos mi

ikanos eimi

elqei

elqein

oitinesnes

oitines

oudnenos

oudenos

touto i

touto

sune[ill]erasure[/ill]

sunestwtas

suntriboun

suntribon

cairete (1) c

cairete (1)

[ill]erasure[/ill]

swma

hdusmon

hduosmon

poisai

poihsai

metauta

meta tauta

ℵ [ill]erasure[/ill]

Of the 20 nonsense words in scribe A’s text, 14 are corrected by the scribe or an early corrector, leaving only six or seven errors.115 Six times a letter or syllable is repeated or added, three of which can definitely be attributed to the occurrence of a line break. Line endings play a role in the It is hard to see who is responsible for the expunction of the repeated syllable -nes at Luke 8:15, folio 77.3 (NT 35), line 3.8. It may well be that one of the early correctors is responsible. 115

Scribes D and A in Luke

227

haplographic spelling of me // tauta for meta tauta; apol // lueis for apolueis; oitines // nes for oitines; and poi // sai for poihsai. However, in some comparable cases no line break is involved: mi (eimi), hdusmon (hduosmon). Additionally, in one case a letter is erased as the scribe decided to start a new paragraph.116 In six cases a letter or syllable is omitted. Three instances are included where the original form is under an erasure and could not be retrieved. In Luke 9:39 the form suntriboun for suntribon is found. In its current form the u is obelised, but it is not improbable that before this it had already been marked by a dot above the letter. This may have been done by the prima manus, which would make the insertion of the u a mere scribal slip and the current text an in-line correction. Gignac indicates that the ou–o interchange is probably an Egyptian feature.117 However, as the interchange ou–o occurs only here and because the form may have never have been intended to be written as such, this is not an argument that can be used in establishing the provenance of Sinaiticus. A variant occurs at Luke 1:69, which we have also seen in the Pauline letters,118 namely, the reduplication of the n of the preposition en. The nu is written as a crossbar above the epsilon but also as the first letter of the new line.

Leaps from the same to the same Luke 12:52, folio 77.7 (NT 39), line 6.37

ℵ diamerismenoi

exemplar diamerismon esontai gar apo tou nun pente en eni oikw diamerismenoi

Only one instance of a leap from the same to the same is found, namely a leap forward in Luke 12:52. Scribe A jumps from diamerismon at the end of verse 51 straight to diamerismenoi in verse 52, skipping nine words. The correction was made by scribe D.119 Luke 9:21. See the discussion on page 102. Gignac I, 213–14. 118 See page 206. 119 The arrows connecting the correction in the lower margin with the place of insertion are clearly scribe D’s. He also changed the final e of line 37 into an o and 116 117

228

Singular Readings and Individual Scribal Patterns

Add and omit verba minora Luke 2:4, folio 76.6 (NT 30), line 3.45 Luke 2:15, folio 76.6 (NT 30), line 5.10 Luke 2:39, folio 76.6 (NT 30), line 8.18 Luke 2:42, folio 76.6 (NT 30), line 8.30 Luke 3:16, folio 76.7 (NT 31), line 4.18 Luke 4:38, folio 76.8 (NT 32), line 3.32 Luke 5:36, folio 77.1 (NT 33), line 1.41 Luke 6:47, folio 77.2 (NT 34), line 1.7 Luke 7:3, folio 77.2 (NT 34), line 2.7 Luke 8:6, folio 77.3 (NT 35), line 1.41 Luke 8:13, folio 77.3 (NT 35), line 2.40 Luke 8:22, folio 77.3 (NT 35), line 4.8 Luke 8:29, folio 77.3 (NT 35), line 5.27 Luke 8:43, folio 77.3 (NT 35), line 7.42 Luke 9:51, folio 77.4 (NT 36), line 8.46 Luke 10:11, folio 77.5 (NT 37), line 3.37 Luke 11:22, folio 77.6 (NT 38), line 3.5 Luke 11:27, folio 77.6 (NT 38), line 3.45 Luke 11:40, folio 77.6 (NT 38), line 5.40

ℵ eis (2) thn polin

exemplar eis (2) polin

poimanes

oi poimanes

galilaian

thn (1) galilaian

ib kai

dwdeka

umas baptisei tou simwnos

autos umas baptisei simwnos

elegen

elegen de kai

twn logwn mou

twn logwn

autos elqwn

elqwn

exhranqh kai

exhranqh

outoi

kai (1) outoi

mia

en mia

diarrhsswn

kai (3) diarrhsswn

kai ouk

ouk

analhmyews geinwskete umis

analhmyews autou (1) ginwskete

autou (1) estin

autou (1)

koilia

h (1) koilia

o afrones

afrones

provided the crossbar of the final nun, leading Tischendorf to read diamerismonnoi as scribe A’s reading.

Scribes D and A in Luke

229

Of the nineteen cases in this category, nine are corrected by the prima manus or an early corrector. The nine omissions include three articles, two pronouns, three conjunctions, and once the preposition en. Of the ten additions, two involve an article and also the exclamation w was added once.120 Three times kai is added, three times a pronoun, and once the implied verb estin. There is no clear tendency in this category to produce a shorter text, as the additions slightly outnumber the omissions.

Harmonisations Harmonisation to the immediate context ℵ Luke 1:13, folio uion sou 76.5 (NT 29), line 4.19

exemplar uion soi

Luke 1:49, folio to eleos 76.5 (NT 29), autou line 8.20

to onoma autou

Luke 2:36, folio chreusasa 76.6 (NT 30), line 7.45 Luke 2:41, folio kat eqos 76.6 (NT 30), line 8.26 Luke 3:17, folio katasbesi 76.7 (NT 31), line 4.27

zhsasa

Luke 4:41, folio 76.8 (NT 32), line 4.9 Luke 7:36, folio 77.2 (NT 34), line 6.46

daimonia pollwn krazontwn katekeito

Luke 8:13, folio logon tou 77.3 (NT 35), qeou line 2.39

120

kat etos katakausei

daimonia pollwn krazonta katekliqh

logon

Luke 11:40, folio 77.6 (NT 38), line 5.40. Written as o.

Previously in the same verse dehsis sou and gunh sou to eleos autou is the clause following on to onoma autou Next verse mentions that Anna is a chra Influence of verse 42 Influence of the second word following asbestw Ending of preceding word taken over Taken from next verse; possible influence from Mark 14:3 o logos tou qeou from verse 11

230

Singular Readings and Individual Scribal Patterns

ℵ Luke 8:29, folio demonia 77.3 (NT 35), line 5.28 Luke 10:30, folio 77.5 (NT 37), line 6.18 Luke 11:28, folio 77.6 (NT 38), line 4.4

exemplar desma

iereicw

ierousalhm

fulassontes ton logon tou qeou

fulassontes

Subject matter of the passage is the exorcism of a daimonia Taken from further down the same verse Repeating the object of the previous participle

Harmonisation to the intermediate context Luke 2:21, folio 76.6 (NT 30), line 6.6 Luke 2:52, folio 76.7 (NT 31), line 1.43 Luke 12:29, folio 77.7 (NT 39), line 3.42

ℵ authn

exemplar auton

qeou

para qew

pihte mhde tw swmati

pihte

cf 2:6 eplhsqh-san ai hmerai tou [infinitive] authn Influence of a similar remark at 2:40 Immediate context v 22, also Matt 6:31. Tischendorf sees hooks around the addition which I fail to see.

Harmonisation to remote context / general usage Luke 5:18, folio 76.8 (NT 32), line 7.17 Luke 7:35, folio 77.2 (NT 34), line 6.40 Luke 8:7, folio 77.3 (NT 35), line 1.48

ℵ anqrwpon epi klinhs beblhblhmenon ergwn

exemplar epi klinhs anqrwpon teknwn

Synoptic parallel Matt 11:19

epnixan

apepnixan

Synoptic parallel Matt 13:7 epnixan there only attested by ℵ D Q f13 565 pc

Synoptic parallel Matt 9:2

Scribes D and A in Luke

231

The various forms of harmonisation number 17 in total. The majority of these readings (11 cases) are due to the influence of the immediate context. In the words of the angel to Zechariah in Luke 1:13, scribe D writes initially uion sou, following h dehsis sou and h gunh sou. This is corrected to uion soi by the scribe himself. In Luke 1:49, to onoma autou is replaced with to eleos autou, borrowed from the first words of the next verse and remaining uncorrected. All the remaining harmonisations are by scribe A. Harmonisations to the immediate context by scribe A include the following cases. The ending of a participle is changed under the influence of the preceding word: kai daimonia apo pollwn krazontwn for krazonta, which changes the meaning of the phrase considerably and which remained uncorrected. In Luke 2:41, Sinaiticus reads kat eqos for kat etos under the influence of the next verse kata to eqos. Influence from words that at the time of the error still needed to be written can also be seen in chreusasa for zhsasa (in the context of the prophetess Anna in the temple, Luke 2:36), katasbesi for katakausei in Luke 3:17, katekeito for katekliqh in Luke 7:36 (uncorrected), and iereicw for ierousalhm in Luke 10:30. It is possible that the form demonia for desma in Luke 8:29 may not be attributable to the occurrence of daimoniou later in the verse but should be seen as influence of the subject matter of the section, which, from verse 27 onwards, is the man having a demon. Twice an addition is found (Luke 8:13 and 11:28) that arose out of harmonisation with the preceding text: logon tou qu+ for logon (uncorrected), and fulassontes ton logon tou qu+ for fulassontes. Of the eleven harmonisations to the immediate context, eight are corrected by the scribe or an early corrector. Three times a singular reading occurs that can be seen as harmonisation to the intermediate context. In Luke 2:21, the text reads eplhsqhsan hmerai oktw tou peritemin authn, where the feminine pronoun authn yields a nonsense meaning and is corrected early on. The reason for the scribe’s error may lie in the similar phrasing found in Luke 2:6: eplhsqhsan ai hmerai tou tekin authn. The second example is found in Luke 2:52, where Sinaiticus has cariti qu+ instead of cariti tw qw+. In verse 40, which is, like verse 52, the closing verse of a section, the phrase caris qu+ occurs. The third

232

Singular Readings and Individual Scribal Patterns

instance is found in Luke 12:29 where the list of things not to worry about is harmonised to verse 22. Two of the three readings are corrected. All three harmonisations to general usage or passages in other books are harmonisations to Matthean parallels. In Luke 5:18, men are carrying anqrwpon epi klinhs beblhblhmenon. The dittography of blh is written over a line division, and the word beblhmenon is taken from Matt 9:2. The normal wording is epi klinhs anqrwpon. In Luke 7:35, Sinaiticus alone reads ergwn for the normal teknwn, borrowed from Matt 11:19. The third harmonisation to another Gospel may be the form epnixan for apepnixan in the Parable of the Seed. In Matt 13:7 the text of Sinaiticus also reads epnixan.121 None of these three harmonisations is corrected.

Editorial readings Luke 1:65, folio 76.6 (NT 30), line 2.1 Luke 2:2, folio 76.6 (NT 30), line 3.32

exemplar dielaleito panta (?) authn apografhn auth h apografh egeneto prwth prwth egeneto ℵ dia

There are two readings in this category, both by scribe A. The rewriting of Luke 2:2 has also been discussed in the previous chapter with regard to paragraph divisions.122 While the exemplar probably read auth h apografh prwth egeneto, the reading of Sinaiticus is authn apografhn egeneto prwth. The case of auth h apografh has been changed, and also the word order of the following two words is different. A new paragraph starts with egeneto. The other reading is a similar misjudgement of the syntax of a sentence. The normal text of Luke 1:65 reads kai egeneto epi pantas fobos tous perioikountas autous kai en olh th oreinh ths ioudaias dielaleito panta ta rhmata tauta. panta is absent in L and quite a few minuscules and cannot be automatically assumed to be present in the exemplar of Sinaiticus. The second kai functions as a separator of two parts of the sentence as it is followed by a finite verb, dielaleito. This basic structure of the sentence is affected by the change made in Sinaiticus, where 121 An alternative explanation is that these are examples of the tendency of scribe A to drop prefixed prepositions, as seen in our study of Psalms. 122 See the discussion in Chapter 3, page 105.

233

Scribes D and A in Luke

dielaleito panta is replaced with the preposition dia. This may reflect a misreading of the preceding sentence where kai en olh th oreinh ths ioudaias is read as an extension of peri pantas tous perioikountas autous. The creation of this singular variant may have been made easier by the fact that dia is the first word of a new column. Both readings are corrected.

Nonsense meanings Luke 2:22, folio 76.6 (NT 30), line 6.14 Luke 2:28, folio 76.6 (NT 30), line 7.2 Luke 3:7, folio 76.7 (NT 31), line 2.41 Luke 4:1, folio 76.7 (NT 31), line 6.48 Luke 4:2, folio 76.7 (NT 31), line 7.5 Luke 7:28, folio 77.2 (NT 34), line 5.39 Luke 8:10, folio 77.3 (NT 35), line 2.17 Luke 10:15, folio 77.5 (NT 37), line 4.13 Luke 12:26, folio 77.7 (NT 39), line 3.21

ℵ parasthsetai

exemplar parasthsai

autos de edexato autos edexato elegon

elegen

hgeto to

hgeto

ouden ouden

ouden

en gennhtai

en gennhtois

tas musthria

ta musthria

ews (2) o

ews (2)

ti dunasqai ti

dunasqe ti

There are nine readings that make no sense in their context but which are grammatical forms in their own right, five of which are corrected. Two of these have a dropped or added syllable or changed verb ending immediately after a line break.123 Another error that can be attributed to the influence of the physical environment of the text is the dittography of ouden at Luke 4:2. This reading on the verso of a folio occurs at a place where the parchment was damaged because the scribe had rubbed something away on the recto. It is not impossible that the torn parchment caused just enough distraction for him to write ouden twice. In three cases an added letter or syllable is found which constitutes a word but does not fit the context. The word de in the construction kai Luke 2:22 parasth- // setai for parasth- // sai; Luke 4:1 hgeto // to for hgeto. 123

234

Singular Readings and Individual Scribal Patterns

autos de seems to suggest that the scribe did not realise that he had already written kai.

Substitutions ℵ Luke 1:26, folio 76.5 (NT ioudaias 29), line 6.7 Luke 1:39, folio 76.5 (NT eporeueto 29), line 7.22 Luke 2:7, folio 76.6 (NT 30), line 4.15 Luke 2:9, folio 76.6 (NT 30), line 4.28 Luke 2:12, folio 76.6 (NT 30), line 4.44 Luke 2:37, folio 76.6 (NT 30), line 8.1 Luke 2:37, folio 76.6 (NT 30), line 8.5 Luke 2:39, folio 76.6 (NT 30), line 8.17 Luke 3:5, folio 76.7 (NT 31), line 2.37 Luke 5:7, folio 76.8 (NT 32), line 5.29 Luke 5:13, folio 76.8 (NT 32), line 6.24 Luke 5:32, folio 77.1 (NT 33), line 1.17 Luke 6:44, folio 77.1 (NT 33), line 8.37 Luke 8:11, folio 77.3 (NT 35), line 2.27 Luke 8:26, folio 77.3 (NT 35), line 4.42 Luke 8:40, folio 77.3 (NT 35), line 7.17 Luke 9:13, folio 77.4 (NT 36), line 3.16 Luke 11:18, folio 77.6 (NT 38), line 2.31

exemplar galilaias / iouda eporeuqh

epi fatnh

en fatnh

epelamyen

perielamyen

epi fatnh

en fatnh

ebdomhkonta tessarwn nhsteias kai dehsin epestreyen

ogdohkonta tessarwn nhsteiais kai dehsesin epestreyan

trociai

traceiai

sunlambanesqe

sullabesqai

tas ceiras

thn ceira

asebeis

amartwlous

blastou

batou

cristou

qeou

katepausan

katepleusan

ton qn+

auton (2)

pleiones artoi pente ti

pleion h artoi pente oti

235

Scribes D and A in Luke

Luke 12:14, folio 77.7 (NT 39), line 1.48 Luke 12:18, folio 77.7 (NT 39), line 2.23 Luke 12:42, folio 77.7 (NT 39), line 5.26

ℵ ef umwn

exemplar ef umas

anoikodomhsw

oikodomhsw

diadounai

diadidonai

In scribe D’s section of Luke, two substitutions are found. In Luke 1:26 galilaias is replaced with ioudaias. Some minuscules read iouda at this point, and though it is conceivable that the current reading is a conflation of iouda with the ending -ias of galilaias because of an incorrectly interpreted correction, it is probably better to see ioudaias as the intended form from the start. Whether the reading in Sinaiticus is derived from galilaias or iouda is difficult to tell. Another substitution in the text by scribe D is the change from the aorist to the imperfect in Luke 1:39, eporeuqh replaced with eporeueto. Neither of these substitutions has been corrected. In scribe A’s text we find 19 substitutions, 10 of which are corrected. Once a change of a verb ending is found, epestreyen for epestreyan (Luke 2:39). Also the prefixed preposition of a verb is changed twice, once an- is added to oikodomhsw, and in Luke 2:9 perielamyen is changed to epelamyen. A change in tense occurs with sunlambanesqe (read: -sqai) replacing sullabesqai (Luke 5:7), and with diadounai for diadidonai (Luke 12:42). Singular readings involving prepositions are epi fatnh for en fatnh in Luke 2:7 and 2:12, and ef umwn for ef umas in Luke 12:14. The case and number of nhsteiais kai dehsesin are changed to nhsteias kai dehsin. In Luke 11:18 oti is replaced with ti, which is not only a very simple change, but also one which was probably corrected immediately. This leaves those substitutions in which one word is replaced with a different word. In Luke 5:32 asebeis takes the place of the related word amartwlwn. Optical confusion may have played a part in the creation of blastou for batou (Luke 6:44). Both are botanical terms (“shoot”; “thorns”) but carry quite different connotations. In Luke 8:11 qu+ is replaced with cu+ in the expression logon tou qu+, but probably already corrected by the prima manus. The substitution of katepleusan with katepausan in Luke 8:26 remains

236

Singular Readings and Individual Scribal Patterns

uncorrected.124 Whether or not the scribe is influenced by anapauesqai (read anapauesqe) in Mark 6:31, cannot be ascertained. However, the similar shapes of the uncial letters a and l may have played a role. The normal text in Luke 8:40 reads prosdokwntes auton. Sinaiticus has instead of auton the words ton qn+. The substitution has the same number of letters and shares the letters -ton-. The age of Anna in Luke 2:37 is given in Sinaiticus as ebdomhkonta tessarwn instead of ogdohkonta tessarwn. There is no obvious explanation for the origin of this variant. If the exemplar had the numerals written as letters the scribe must have read od+ for pd+.

Transpositions Luke 5:5, folio 76.8 (NT 32), line 5.14 Luke 6:26, folio 77.1 (NT 33), line 5.46 Luke 9:28, folio 77.4 (NT 36), line 5.32

ℵ eipen simwn

oi anqrwpoi pantes proseucasqai kai en tw proseuxesqai Luke 9:41, folio 77.4 (NT meq umwn esomai 36), line 7.21 umwn o pathr Luke 12:32, folio 77.7 (NT 39), line 4.9

exemplar simwn eipen pantes oi anqrwpoi proseuxasqai kai egeneto en tw proseucesqai esomai meq umwn o pathr umwn

Five transpositions occur, all in the text by scribe A and none are corrected. One of these involves two words (Luke 5:5 eipen simwn), three involve three words (Luke 6:26 oi anqrwpoi pantes; Luke 9:41 meq umwn esomai; Luke 12:32 umwn o pathr). One transposition involves the swapping of two closely related forms, proseuxasqai and proseucesqai in Luke 9:28 and 29. The two infinitives are, in Sinaiticus, only separated by kai en tw and have been interchanged.125 kataplew “to sail off,” katapauw “to stop, rest.” The apparatus of NA27 suggests that proseuxasqai was written in verse 29 under the influence of verse 28. This may have been the case of the other cited witnesses for this reading but not for Sinaiticus. The word order in verse 29 is further affected by the various placements of egeneto. Sinaiticus agrees with 1071 in placing egeneto after autou (1), though 1071 also has, according to 124 125

237

Scribes D and A in Luke

Add and omit words and clauses ℵ Luke 2:27, folio 76.6 (NT to paidion 30), line 6.46 Luke 2:37, folio 76.6 (NT ek tou ierou 30), line 8.4 Luke 3:1, folio 76.7 (NT pontiou pilatou 31), line 1.48 Luke 4:1, folio 76.7 (NT omitted 31), line 6.47 Luke 4:42, folio 76.8 (NT erhmon 32), line 4.18 Luke 5:1, folio 76.8 (NT gennhsaret 32), line 4.40 Luke 5:2, folio 76.8 (NT ploia 32), line 4.42 Luke 5:2, folio 76.8 (NT autwn apobantes 32), line 4.44 Luke 5:14, folio 76.8 (NT omitted 32), line 6.32 Luke 6:14, folio 77.1 (NT omitted 33), line 4.21 Luke 7:11, folio 77.2 (NT nain 34), line 3.18 Luke 8:30, folio 77.3 (NT omitted 35), line 5.32 Luke 8:36, folio 77.3 (NT de legontes 35), line 6.33 Luke 8:47, folio 77.3 (NT omitted 35), line 8.17 Luke 8:55, folio 77.4 (NT 36), line 1.11 Luke 9:7, folio 77.4 (NT 36), line 2.10 Luke 9:43, folio 77.4 (NT 36), line 7.41 Luke 10:42, folio 77.5 (NT 37), line 7.48

omitted hrwdhs

omitted

exemplar to paidion ihsoun tou ierou pontiou pilatou ths ioudaias upestreyen erhmon topon limnhn gennhsaret duo ploia ap autwn apobantes deixon seauton tw ierei kai kai barqolomaion kaloumenhn nain o ihsous de idousa de h gunh oti ouk elaqen tremousa hlqen kai anesth paracrhma hrwdhs o tetrarchs eipen

oligwn de estin h oligwn de creia enos estin h enos

IGNTP, egeneto at the beginning of this verse. The apparatus to this verse in IGNTP is structured in a rather confusing way.

238

Singular Readings and Individual Scribal Patterns

ℵ Luke 11:1, folio 77.5 (NT kaqws 37), line 8.14 Luke 11:2, folio 77.5 (NT ouranw outw 37), line 8.25 Luke 12:8, folio 77.7 (NT tou q)u 39), line 1.12

exemplar kaqws kai iwannhs ouranw aggelwn tou qeou

Of the 21 instances in this group of omitted or added words or clauses, nine are corrected. Three times a word is added, 18 times a word or clause is omitted. All the instances occur in scribe A’s text. The first addition is found in Luke 2:37 where ek is inserted before the article in the phrase h ou afistato ek tou ierou. In Luke 8:36, the participle legontes is inserted: aphggeilan de legontes autois. The third example is found in the Lord’s Prayer in Luke 11:2. According to the available evidence, Sinaiticus is the only manuscript that reads outw in the phrase ws en ouranw kai outw epi ghs.126 As regards the omissions, in one case a preposition is omitted: apo is left out in Luke 5:2: oi de aleeis ap autwn apobantes. In Luke 4:1, the finite verb upestreyen is initially omitted, and likewise eipen is left out in Luke 9:43, both omissions resulting in a strained syntax. In Luke 4:41, the second word of erhmon topon is left out. In the latter case the meaning is not affected nor is there an obvious gap left in the sentence. The same is true for the following nine omissions (the omitted words are underlined) to paidion ihsoun; pontiou pilatou ths ioudaias; limnhn gennhsaret; duo ploia; kai iwannhn kai filippon kai barqolomaion; kaloumenhn nain; ephrwthsen de auton o is+; hrwdhs o tetrarchs; and aggelwn tou qu+.127 Origen deals with this phrase extensively in De Oratione but does not know this addition in the Lucan version of the Lord’s prayer (26.1). In 26.6 he cites our phrase in the following form: ws en ouranw outws kai epi ghs. In Koetschau’s edition, the word outws is punctuated as not being part of the citation but as part of the framing phrase by Origen. Given the numerous other citations of these words in the same context this seems highly probable. Origen cannot, therefore, be regarded as a witness to the text of Sinaiticus. 127 Regarding the omission of aggelwn in Luke 12:8, it should be noted that NA27 is wrong regarding the omission of this word in the parallel expression 126

Scribes D and A in Luke

239

Sinaiticus has a unique reading in Luke 10:42 in the midst of a confused textual tradition. The readings that come closest to the version in Sinaiticus are oligwn de creia estin h enos (only attested by B) and oligwn de estin creia h enos (P3 Cc2 L f1 33). Sinaiticus has oligwn de estin h enos, omitting the necessary creia. On the basis of the evidence for these two readings, it is hard to decide which is the reading of the exemplar. However, the physical position of the omission may be helpful here. estin is the last word of a column, the next starts with h enos. As we have seen already that singular readings were triggered by line and column breaks, it is likely that the omission of creia occurs for the same reason. Therefore it seems more plausible that in the exemplar of Sinaiticus creia followed estin.128 Three times a larger stretch of text is omitted, though in none of these cases can this be considered a leap from the same to the same. In Luke 5:14 the words deixon seauton tw ierei kai are left out. Though this omission is probably a mere oversight and does not distort the flow of the sentence, it does happen in a syntactically interesting context. Verse 14 opens with a verb of communication, parhggeilen, and is followed by indirect discourse. Luke then continues with the imperative deixon turning to direct discourse. In many manuscripts this is followed by a second imperative kai prosenegke, but a considerable number of manuscripts have instead an infinitive, kai prosenegkai, which is dependent upon parhggeilan. Sinaiticus has also the spelling prosenegkai, but this could easily be a common itacism. Interestingly though, as the whole phrase containing the imperative deixon has dropped out and the imperative/infinitive is spelled as if it is the infinitive, the version of Sinaiticus does not leave the indirect discourse: parhggeilen autw mhdeni eipein alla prosenegkai. Yet it is hardly likely that this reworking of the syntax was intentional; the personal pronoun of tou kaqarismou is left in the second person singular. The other two large omissions are found in Luke 8:47 and 8:55 but they do not affect the flow of the text. Luke 8:55

in the next verse. Influence from related expressions (general context) cannot be excluded. 128 Alternatively one can think of a misinterpreted correction in the exemplar in which creia was transposed from before estin to after estin (or vice versa). The simpler explanation is preferred here.

240

Singular Readings and Individual Scribal Patterns

might be regarded as a leap from the same to the same, as both the omitted clause and the next clause start with kai.

Rewritings Luke 2:3, folio 76.6 (NT 30), line 3.38

Luke 8:47, folio 77.3 (NT 35), line 8.18

ℵ ekastos apografesqe eis thn eautwn polin dihggeilen

exemplar pantes apografesqai ekastos eis thn eautou polin di hn aitian hyato autou aphggeilen

Two readings show a more complicated rewriting, involving more than one change to the text. In Luke 2:3 the normal text reads kai eporeuonto pantes apografesqai ekastos eis thn eautou polin. The reading in Sinaiticus has three differences: kai eporeuonto ekastos apografesqe (read -sqai) eis thn eautwn polin. pantes has been omitted, its place is taken by ekastos129 and the singular eautou is made into a plural. This reading is left uncorrected. The second rewriting is found in Luke 8:47. Under “Add and omit words and clauses” we have already mentioned the major omission of the first half of this verse, yet in the second half of the verse another shortening of the text occurs. The standard text reads kai prospesousa autw di hn aitian hyato autou aphggeilen which is shortened in Sinaiticus to kai prospesousa autw dihggeilen. It is possible to see this reading as a leap forward from the same to the same, namely from di hn- to the similar sounding dihng-/dihgg-. The result of this change is an omission and a substitution. This reading also remains uncorrected. Conclusions Before correction, the six columns of text by scribe D contain six singular readings, which would translate to 80 singular readings per 10 folios. Compared to scribe A (126 singular readings per 10 folios), this is a lower number, but the difference is not excessive. Just as in the replacement folios of the Pauline corpus, a difference in the frequency of singular readings is The constructio ad sensum of having the singular distributive ekastos with a plural verb is widespread in the NT and the papyri. See BDF, p. 74. 129

Conclusions

241

more likely to reflect a difference between the two scribes than a difference in the quality of the exemplar. However, one can never totally exclude the possibility that a different exemplar was used for the replacement folios. If we assume that both scribes use the same exemplar, the copy of the exemplar made by scribe D is more precise and accurate than the text by scribe A. In other words, the text of scribe D approaches the exemplar more nearly than does that of scribe A. This gives us the opportunity to describe the scribal habits of scribe A in Luke with much more confidence than if we did not have the “benchmark text” of scribe D. Scribe A creates more nonsense words than scribe D and has a clear tendency to add and omit verba minora. In Luke the numbers of such omissions and additions are about equal. When it comes to the more substantial words and phrases, the tendency to omit rather than add is clearly present. Harmonisation to the immediate context is a recurrent habit; the influence from the other Gospels is also noticeable. There is no obvious tendency to change text because of general usage or attempts to improve the text in a conscious way. Substitutions occur regularly, though most substitutions affect the declension or inflection of words. There are quite a few larger omissions of whole phrases and words in apposition. We have not seen this in the same measure elsewhere in scribe A’s work, and therefore these omissions may be part of the inherited text (though no comparable cases are found in scribe D’s section of Luke).

CONCLUSIONS On using singular readings to study scribal habits Now that we have analysed a large number of singular readings, can we say anything more about the method of studying singular readings to learn about scribal habits? First, the categories we used were largely dependent on Colwell and Royse. Though these categories may have served to ask certain questions of the evidence, not each of these categories was particularly helpful. The categories can be placed on a spectrum ranging from mechanical errors to editorial changes. Those readings that lean more towards the editorial side were, in our classification, subdivided into misreading of syntax, misinterpreted corrections, smoothing of the text, and readings affected by interfering meaning. However, the “major rewritings” and “harmonisations to general usage” are closely related to these groups. Given the relatively low frequency of readings classifiable as “editorial,” the various categories could perhaps be collapsed into one group.

242

Singular Readings and Individual Scribal Patterns

Secondly, the basic premise of Colwell’s approach is to filter out first the orthographic readings and nonsense words, and then to deal with those readings which have common, scribal explanations, such as leaps from the same to same, add and omit verba minora, and harmonisations. To include the addition and omission of verba minora at this stage is completely justified. Variants in this category are numerous and seldom affect the overall meaning of the text. In our study we could see a considerable laxity by scribe A on this point and in Luke there was even a stark contrast between the ratio of omissions and additions of verba minora and the ratio of omissions and additions of more substantial words and clauses. Thirdly, the special feature of Sinaiticus in which the work of two scribes exists side by side has enabled us to confirm that the method of singular readings indeed delivers what it claims: in the case of Sinaiticus we have a window into the actual habits of the individual scribe. The differences between the two sections of Psalms and the difference in the frequency of singular readings between the replacement folios in the NT and the surrounding text are best explained by reference to the scribes who made the copy. We are even able to gain more insight into one of the two uncertainties surrounding the use of singular readings: the number of readings that are falsely excluded. We noted that among the corrected and scribe-created readings of 1 Chronicles, 13 out of 90 readings are found in other manuscripts as well. Given the fact that we do not have many manuscripts of 1 Chronicles, it is likely that some of the other readings would show up elsewhere if we had more manuscripts. If we extrapolate this to the New Testament, where we have more manuscripts, it follows that the scribes must have made considerably more errors in copying the exemplar than actually show up as singular readings. If one would test and compare the scribal habits of modern scribes in laboratory circumstances, one would ask both scribes to copy the same text and compare their work afterwards. In Sinaiticus we may have a situation that is as close to this as we can get for an ancient manuscript: here two scribes work on different parts of the same text, where the work of one of them (scribe D) is done under two different sets of circumstances. In Psalms, scribe D functioned as a normal scribe, and, in Luke and Paul, the same scribe replaced an already existing folio in the function of corrector. The individual scribal patterns for scribes D and A Though scribe D did contribute to the Pauline and Lucan text of Sinaiticus, the total length of his contribution is small. It is therefore difficult to make a meaningful comparison using the replacement folios. When these two

243

Conclusions

contributions in the NT are compared with the much larger section of Psalms copied by scribe D, it becomes clear that both replacement sheets fall within the normal range of scribe D’s habits. Number of singular readings per 10 folios Orthography Nonsense word forms Leaps a/o verba minora Harmonisation - immediate context - intermediate context - general usage Editorial readings - interfering meaning - misreading syntax - misunderstood corrections - deliberate improvements Nonsense meanings Substitutions Transpositions a/o words or clauses More complex rewritings Total:

D - Psalms

D - Luke

7.2 6.0 6.0 7.2 11.3 5.3 3.4 2.6 1.1 0.8

26.7

0.4 4.5 12.5 3.0 6.0 1.9 66.8

D - Paul

5.2 26.7 26.7

26.7

15.6 15.6

80.0

36.5

Based on the study of Psalms, the following remarks about scribe D can be made: The two vices to which he is most particularly prone are substitution and harmonisation. These are followed by the addition and omission of both verba minora and of more substantial words and clauses, leaps, and the creation of nonsense words (though very often these are corrected immediately). Though the number of orthographic peculiarities looks high, these are restricted to a few words only. There are very few transpositions or complex rewritings, and the number of replacements is always lower than that of scribe A. The work of scribe A can be compared over four sections: 1 Chronicles, Psalms, Paul, and Luke. The results are found below: Number of singular readings per 10 folios Orthography

A - 1 Chr ATotal130 Uncorr. Corr. Psalms SR 161.9 121.5 42.5 3.9

APaul

ALuke

5.5

11.1

The column “Total” adds up the uncorrected singular readings of 1 Chr and the corrected readings. 130

244

Singular Readings and Individual Scribal Patterns

Number of singular readings per 10 folios Nonsense word forms Leaps a/o verba minora Harmonisation - immediate context - intermediate context - general usage Editorial readings - interfering meaning - misreading syntax - misunderstood corrections - deliberate improvements Nonsense meanings Substitutions Transpositions a/o words or clauses More complex rewritings Total:

A - 1 Chr Total130 Uncorr. Corr. SR 20.2 4.0 14.2 6.1 4.0 2.0 58.7 32.4 26.3 56.7 26.3 30.4 46.6 24.3 22.3

APsalms

APaul

ALuke

16.2 9.1 28.6 17.5 14.9 1.9 0.6 3.9 0.6 2.6 0.6

12.4 4.1 10.4 9.0 4.8 1.4 2.8 2.1

20.3 1.0 19.3 15.2 9.1 3.0 3.0 2.0

10.1 36.4 2.0 26.3

2.0 18.2 16.2

8.1 18.2 2.0 10.1

8.1 28.3 68.8 12.1 32.4

2.0 18.2 52.6 10.1 10.1

6.1 10.1 16.2 2.0 22.3

9.7 40.3 3.9 8.4

479.8

297.6

182.2

140.9

0.7 1.4 7.6 12.4 4.8 11.7 2.1 82.1

2.0 9.1 19.3 5.1 21.3 2.0 125.6

The numbers for 1 Chronicles are extremely high when compared to the other three sections. However, the section of 1 Chronicles was studied mainly because the corrections provide a catalogue of scribe-created readings. It is remarkable how the number of corrected readings in 1 Chronicles form an approximation of the singular readings in the other three sections (with the exception of orthography). The results from the book of Psalms are reasonably close to the two sections from the New Testament and there is no apparent reason why these should not be used to establish the profile of scribal habits for scribe A. A consistent characteristic of scribe A is the frequent creation of nonsense words. Also a word is often substituted for another form of the same word or another word altogether. This is followed by a tendency to add or omit verba minora and words or clauses; in both categories the omissions tend to outnumber the additions. Interestingly, many of the omissions of a word or phrase leave a perfectly normal sense behind. This means that these omissions are not to be explained by mere oversight but rather took place in the memory of the scribe. The scribe is likewise inclined to harmonise the text to its immediate environment. Though leaps do occur, these are not particularly frequent in the prose texts, the higher number in Psalms being attributed to the omission of whole poetic lines.

Conclusions

245

Transpositions do not happen often. When they do, they generally do not involve a great number of words: most concern only two clauses that change position. The absence of orthographic variants in Psalms (besides the itacisms discussed in Chapter 3) stands out; in the prose sections scribe A tends to have a higher number of such singular readings. Also we have seen a number of orthographic variants involving the absence of an augment in compound verbs.131 The more complex rewritings happen only occasionally. Scribe A frequently creates new readings in his copying. Quite a few of these are obvious creations by the scribe, though there is also a group of scribe-created readings that would survive unnoticed by a reader or by a next scribe. A few generations of scribes with the same scribal habits as scribe A would create a text that deviates in many of its details from the exemplar of scribe A. The nature of some of the harmonisations to remote contexts shows that the scribe had a good knowledge of the New Testament. As to the question asked in the introduction, which was whether individual copying habits exist, the answer is affirmative. Scribe A and D show different scribal habits and individual characteristics do appear in the actual handling of the text itself. In the case of Sinaiticus this means that it is not enough to take only the characteristic patterns of the document as a whole into account—it is also necessary to be aware of which scribe is responsible for which text. On scribal habits in general Throughout this chapter the influence of line and column endings has been noted in the discussion of readings. This influence is clearly noticeable in cases where a syllable is repeated before and after the line break. Line and column breaks also play a role in a number of omissions. However, a break is nothing more than a contributing factor which increases the chances of an error; there is not a single category of readings that is limited to line breaks which does not also occur within a line. We have also encountered an example where the bad physical condition of the parchment is likely to have contributed to the creation of a single reading.132 Harmonisation to the immediate and intermediate context does occur frequently. In Sinaiticus it often happens that the text is harmonised to words and clauses further down in the text. The text that follows 131 132

See above pages 149, 172, and 224. Luke 4:2, folio 76.7 (NT 31), line 7.5. See above page 233.

246

Singular Readings and Individual Scribal Patterns

sometimes exerts an influence on the text that is being written. Whether this happens because the scribe’s eye happens to fall on text further down or because the scribe has read ahead is impossible to say.This is an important phenomenon and should be included in the standard canons of New Testament textual criticism. As mentioned in the introduction to this chapter, Royse presented as one of his main findings that scribes omit rather than add words. This was confirmed by Head’s work on the fragmentary papyri. Now we can add the result of our analysis to these studies. First, in Psalms the majority of leaps from the same to the same consists of the omission of a poetic line. This occurs in the work of both scribes. In Luke there is only one clear-cut example of a leap forward, and in the Pauline section the number of forward leaps is equal to the backward leaps. If we concentrate only on the prose sections we find 3 leaps backward and 4 leaps forward. It is only in the leaps forward that text is left out. Secondly, the variants surrounding verba minora show a clear tendency to omit rather than to add. In Psalms scribe D has 4 additions against 15 omissions, for scribe A the figures are 12 and 31. In Paul omissions are again more frequent, but in Luke there is one addition more than there are omissions. Thirdly, in the omissions of words or clauses in Luke, we find only 3 additions against 18 omissions with similar numbers in the other two sections of Sinaiticus we examined. It will be clear from this that the scribal tendency to omit rather than to add is in Sinaiticus similar to that found by Royse in the papyri. In Luke an exception exists with regard to the group of verba minora, but not when it comes to the more substantial words and clauses. In the category of the major rewritings, all six cases in Psalms are shorter than the reading of the exemplar. In the Pauline corpus two rewritings are of a similar length and one is shorter, and the two rewritings in Luke are also both shorter. Royse wanted to “penetrate the fourth-century barrier” in order to study the scribal habits of scribes from a period in which the scribal tradition was less rigorously controlled. It appears that, in Sinaiticus, we have an example of a post-300 AD manuscript that behaves as the pre-300 papyri. Sinaiticus may be evidence for the non-existence of the “fourthcentury barrier” as scribe A clearly stands in the tradition of the lessrigorous approach to careful copying. He works side-by-side with scribe D who is arguably more precise. One may wonder, therefore, whether it is correct to split the history of textual transmission up into rigorously and non-rigorously controlled periods.

5 FINAL REFLECTIONS

SCRIBAL BEHAVIOUR In the previous three chapters, a variety of scribal habits have been studied. In Chapter 2, we looked at the interaction between scribes A and D based on a close examination of the constitutive elements of the codex such as the quire, page, and column. These two scribes have a close working relation. In three books, both write a part of the main text (4 Maccabees, Psalms, Revelation), and scribe D often corrects the work of scribe A. In the NT, scribe D three times replaces a sheet originally written by scribe A. Scribe B, on the other hand, produces his sections independently from the other two scribes. He does all the work on his text by himself, including the corrections and the additional apparatus such as the tituli. Clear signs of the text being “stretched” in the closing stages of Judith lead us to conclude that this book had to be fitted to the next book. The opposite phenomenon, that of squeezing the text in, was found in 4 Maccabees, which likewise had to be fitted to a next section or was for other reasons required to be contained within a single quire. The division of the various tasks seems to have varied from place to place and was not standardised. In Chapter 3, we studied numerous instances of differences in scribal behaviour between the three scribes. None of the three scribes has the same policy in using nomina sacra, they all differ in the frequency of their preferred contractions as well as in some of the actual forms of the contraction. Reverential motives do not seem to play an obvious role for the scribes, most nomina sacra are used in both profane and sacral contexts. An exception is the way in which scribe D uses the nomen sacrum for kurios. Here a distinction is made between the contracted form, used mostly for the Godhead, and the plene form, which is used for a profane referent. Also, most uncontracted forms of ihsous do not refer to Jesus Christ but to the high priest Joshua (in scribe B’s text) and to Jesus Justus (in scribe A’s text). Another difference between the scribes showed up in the use and function of some of the ligatures. For scribe D, the 247

248

Final Reflections

combination of letters sharing a vertical stroke has a clear ornamental value and is used considerably more often in Psalms than in prose texts. The same is true for the mou-ligature for both scribe D and A. The characteristic patterns of the most common orthographic changes (often called itacisms) are so clear that they could almost serve in their own right to identify the three scribes. However, we encountered an unexplained change in scribe A’s normal pattern in most of Matthew. If the ability to spell correctly is indicative of the quality of copying, then scribe D is the best scribe. Scribe B has not a firm grasp of standard orthography. In our discussion of paragraphing, we assumed that it was a feature that should be studied in Sinaiticus on the scribal level rather than on that of an inherited tradition and this proved indeed a fruitful approach. Again, differences in the quality and quantity of paragraphs were found between the scribes, though also within the work of a single scribe the frequency of paragraphs can swing significantly. At places where the text needs stretching out, the number of paragraph breaks increases significantly. The Eusebian apparatus is added by scribes A and D, but no difference was found between the work of these scribes. A marked difference exists between the quality of the Eusebian apparatus from gospel to gospel. Three other numbering systems, found in Acts, Ecclesiastes, and Song of Songs, are all by scribe A in text copied by himself. None of the systems is inserted in its entirety. Also a unique list of kephalia, unparalleled in any other manuscript, is added by scribe A in the upper margin of Acts. In Chapter 4 we discussed the individual patterns of singular readings in sections of 1 Chronicles, Psalms, the Pauline corpus, and Luke. As the last three sections were chosen because of the occurrence of contributions by scribe D and A side by side, we could establish characteristic scribal habits for both these scribes. Clear differences exists between these scribes in the inclination to create nonsense forms, in making leaps from the same to the same, in substituting words, and in adding or omitting verba minora (small words such as conjunctions and pronouns). Overall, scribe D is the more precise scribe of the two, but the difference from scribe A is not similar in all categories. In Psalms, for example, scribe D has a higher frequency of leaps. Scribe D was pronounced a better scribe than scribe A by Milne and Skeat: we have sought to demonstrate how much better he was. Our studies in the scribal habits have probed into the complex issue of the numerous differences in actual scribal practice and ability between the three scribes. Building on the work of Milne and Skeat, we have taken the study of Sinaiticus one step further and have drawn up a more complete and

Scribal Behaviour

249

detailed profile of each scribe, particularly of scribes D and A. Because of the specific features of Sinaiticus, where two scribes worked in close association (one even replacing work by the other), we could extend the profile of the scribes into the way in which they actually handled the text. This last category is, of course, the most significant for the textual criticism of the Greek Bible and adds a complication to the citation of Sinaiticus in any critical apparatus. In Sinaiticus there is not homogeneity in the quality of copying; different scribes produce a demonstrably different quality of text. It is even not enough to simply refer to the quality of the text in a certain book of Sinaiticus: rather, one must go one step further because the identity of the scribe is important. In books such as 4 Maccabees and Psalms, it makes a real difference whether a reading is attested by scribe D or by scribe A; the former scribe is more likely to have made a faithful copy of his exemplar. Likewise, the same difference exists for parts of the New Testament, where the individual witness of Sinaiticus has a different weight in the work by scribe D, the six replacement folios and the opening verses of Revelation, than in the work by scribe A. We have demonstrated that scribe A creates a considerable number of readings whilst copying the text. We only looked at the singular readings, but it is likely that a considerable number of scribe-created readings arose independently in other manuscripts. Besides the obvious differences between the scribes, we are faced again and again with the complexity of the whole manuscript. Three scribes worked together, but certainly not in a straightforward and uniform manner. Procedures for the final editing of the manuscript changed along the way and different scribes executed the various tasks in different parts of the codex. However, even before the final editing the scribes were faced with problems when they struggled to fit the various contributions to the actual text of the manuscript together. A further complicating element is that not every part of the codex received equal attention. The book of Psalms, for example, was written with attention for the aesthetic aspect, as evidenced by the deliberate choice of red ink for the headings and numbering, but was not properly corrected. Romans receives a very dense paragraphing, while in Revelation the paragraphs are sparse and inconsistent. The three scribes have their own specific habits, but nevertheless display considerable variation within their own work. A final notion that comes out of this study of scribal behaviour is the inter-relatedness of the various phenomena in the codex. Quire boundaries, change of scribal hand, line length, density of paragraphing, and the use of nomina sacra can all constitute the cumulative evidence for problems the

250

Final Reflections

scribes faced in the production of the codex. In our study we frequently encountered examples of the influence of line and column breaks on the occurrence of scribal slips, and once a tear in the parchment caused enough distraction that the scribe produced a dittography. In the case of Sinaiticus, knowledge of the manuscript for text-critical purposes includes knowledge of its physical composition and of the interplay of the scribes with the material they were writing on as well as their interaction with one another.

THE CODEX AND THE EXEMPLAR Occasionally our studies have given us a glimpse into the nature of the exemplar, but these insights are not enough to create anything like a comprehensive picture. On the basis of the frequent omissions of whole lines we suggested that the exemplar of the book of Psalms was most likely written in stichoi lines, just as Sinaiticus is. The exemplar that provided the text for 2 Esdras must also have contained 1 Chronicles, and likely 2 Chronicles and 1 Esdras. The two books of the Maccabees that are included in Sinaiticus, 1 and 4 Maccabees, seem to have been copied from (a) source(s) that did not contain the text of 2 and/or 3 Maccabees. In the NT, we have suggested that the Eusebian apparatus may not have been copied from a single source. The nature of the errors in the Eusebian apparatus differs from gospel to gospel, which suggests that different exemplars were used. Admittedly, this may have occurred further back in the ancestry of the source that provided the Eusebian apparatus.

DICTATION THEORY Having studied the manuscript in great detail, we are now in a position to consider the issue of dictation again. If Sinaiticus was produced by dictation, the setting of its production must have been one in which a reader read the text aloud so it could be simultaneously written down by a number of copyists. In Chapter 2 we have looked at signs of squeezing and stretching of the text. It seems unlikely that a scribe who has to follow the tempo that a reader sets and cannot browse a few folios ahead to see how long the text is (because he has no exemplar in front of him), would still have time to think about the total length of his text and the amount of space he has left in the quire. However, we saw this happening in 4 Maccabees, where scribe D and scribe A worked together in a concerted effort to keep the book within the limits of a quire, and in 1 Maccabees where the quire length was adjusted because of considerations of space. Likewise, the dictation theory does not fit well with the case of a scribe who starts writing a new book while his colleague is still working on the previous book, as happens in the

Dictation Theory

251

transition from Judith to 1 Maccabees. Also, in Psalms, a difference between the two scribes exists in the number of erasures and subsequent rewritings. However, for dictation to work, all the scribes must be working at a reasonable pace; the speed of the whole group is determined by the slowest scribe. In the dictation scenario, the existence of erasures and rewritings can only be explained if these were done after the actual dictation session. If, however, a scribe copies on his own, he can take the time to stop and make corrections in the middle of a word. Another problem that cannot be explained by the dictation theory is the changed orthographic pattern of scribe A in the first two quires of Matthew. Is it possible within the dictation theory to explain why the scribe deteriorates in some departments and improves in others? The main argument in favour of dictation is the distinctive orthography of the three scribes. If a scribe is not copying from a physical exemplar lying in front of him but instead listens to a reader, orthographic variants are much more likely to creep in: the exemplar does not provide the scribe with a visual impression of the Greek text. We could add to this the apparently uncontrolled insertion of paragraph breaks, which seems also to display individual habits of the scribe rather than a more or less homogenous pattern inherited from the exemplar. Likewise, the treatment of nomina sacra differs between two scribes working on the same book. All three of these phenomena could be used to argue for a considerable distance between the actual shaping of the text by the scribe and the way the text looks in the exemplar. However, these arguments do not outweigh the earlier objections against dictation, especially since the apparent distance between exemplar and copy is also explained in the model of the copying process developed by Dain and used by Junack.1 In this model the copying process is divided into four stages: reading the text, retention of the text in memory, internal dictation, and the actual movement of the hand. Orthographic variants in which vowels are replaced by similar sounding vowels will arise in the third stage and will be particularly numerous if the scribe has no notion of “pure” orthography (either through indifference or lack of education). Milne and Skeat added also a textual argument in support of dictation. At 1 Macc 5:20 the text reads h_ § h_ ‚i-_ (interpreted by Milne and Skeat as h ex h triscilioi), supposedly the result of a reader exclaiming “either six or three thousand.”2 Milne and Skeat mention the strokes above the 1 2

See Chapter 1, page 23. See above Chapter 1, page 22. Milne-Skeat, 57.

252

Final Reflections

first, third and fourth letter and interpret this stroke, which normally would indicate a numeral, as reflecting the confusion of the scribe. They do not explain, however, why the scribe would be confused by the exclamation of the reader. The reading as it stands, with the crossbars, is a clear nonsense reading, but it is not necessarily a reflection of how a reader tried to correct himself and confused a scribe in the process. Moreover, the numeral used for three thousand, occurs, as far as I can tell, only a couple of times in 1 Maccabees, and never outside this book, which suggests that in this instance we have an argument for a visual link between Sinaiticus and the exemplar. The irregularities found in the work of scribes A and D are evidence of the copying procedure that was used, and it does not seem that this procedure leaves room for the suggestion that Sinaiticus was written by dictation.

DATE AND PROVENANCE Past discussion on the date and provenance has often been accompanied by a fair amount of speculation which tried to fit the manuscript into a known context. No certainty can be achieved on the issues of date and provenance, but the weight of probabilities may lend support for or disprove certain contexts. In Chapter 1 we have reviewed the case T.C. Skeat has made for a Caesarean origin of Sinaiticus. The positive, “incontrovertible” arguments from within the codex come from errors in the transcription or insertion of certain place names: antipatrida for patrida in Matt 13:54; kaisareias for samareias in Acts 8:5, and ippon for ioppon in 1 Macc 14:5, all of which are said to point to local knowledge. However, if we have a second look at these names, they turn out to be far from convincing. All three readings come from sections written by scribe A, and we have only to remember the many variants in proper nouns in 1 Chronicles to realise that an argument based on the copying technique of this scribe is hazardous. The first variant is given most attention by Skeat. The text in Sinaiticus looks thus: Matt 13:54, folio 73.8 (NT 8), line 6.29–30 kai elqwn eis th ( aªnªtªi°patrida autou

The first four letters are expuncted and it is notoriously difficult to ascribe such correction to a specific corrector. It is certain that none of the C-correctors is responsible, as they use more than one dot per letter. We

Date and Provenance

253

cannot be sure, however, who of the earliest correctors was responsible: it may have been the prima manus, scribe A. Neither can we be certain if the reading antipatrida was the intended reading. A possibility is that we are dealing with an abandoned substitution and that the scribe intended to write antiperan but realised and corrected his error before finishing the word. But even if this is not the case, it is difficult to see this reading as support for Caesarea as the place of origin for Sinaiticus: at best it tells us only something about the scribe. Moreover, Antipatris is not only 45 km. removed from Caesarea, but also about the same distance away from another centre of manuscript production, Jerusalem.3 The second argument involving a place name, samarias replaced by kaisarias, is equally problematic as an argument for Caesarea as the place of origin of Sinaiticus, as there is a better explanation for this change. Apart from the close similarity in shape of the two names, the context of this replacement provides a clue: the subject of this section is Philip, the deacon, who, further down in Acts, is seen as living in Caesarea. It is much more in line with the demonstrated habits of the scribe to explain this substitution by means of harmonisation to the intermediate context than anything else. The objection that the scribe has not yet copied that part of Acts does not hold. We have not only seen more examples of influence from still-to-be-copied text, but also the scribe clearly wrote from a background in which he had knowledge of the biblical text. Thus, it appears that there are no real arguments from within the codex that argue in favour of a Caesarean origin. But are there any arguments against it? A possible argument can be construed on the basis of the Eusebian apparatus. Though its presence may initially look like a strong argument in favour of the Caesarean hypothesis because of Eusebius’s connection with the city, it may on closer inspection actually point away from Caesarea. The Eusebian apparatus that has been written into Sinaiticus is a version that has suffered already during its transmission history. One would assume that, had Sinaiticus been copied in Caesarea, the system would have been included in a purer form than it is now. The state of the Eusebian apparatus suggests a certain distance from its source. More speculative arguments could include the difference in abilities of the three scribes. The work of scribe A and scribe B is not of a very high quality. One might expect that a scriptorium such as Caesarea, or any other major centre, would guard their reputation better than to produce a Bible of Compare the journey of Paul from Jerusalem to Caesarea in Acts 23:31–33, during which the whole company stopped over at Antipatris. 3

254

Final Reflections

this quality. However, this argument is based on how we envisage a fourthcentury scriptorium, which may be completely beside the historical reality. Another argument pointing away from Caesarea, and likewise from other major centres, is that the codicology of Sinaiticus shows various irregularities in the quire make-up as well as a shift in the way quires were initially fastened together for pricking and ruling. These irregularities betray a nonstandardised production process and give at times an impression of inexperience. But again, this argument is based on how we expect a scriptorium would have functioned and may not have much force at all. A more general argument may be found in the possible implications of the sheer size of the folio and the phenomenon of producing a complete Bible. In a survey of the earliest Latin biblical manuscripts, Patrick McGurk showed that the first large and complete Bibles appeared relatively late, from the sixth century onwards.4 That a large Greek Bible appeared so much earlier might suggest that it was made in an environment which was not very influential on Latin Bible production. This argument equally applies to Vaticanus, Ephraemi Syri Rescriptus, and Alexandrinus, the other complete Greek Bibles from the fourth and fifth centuries. But again, the available evidence does not allow us to say much more. The only firm evidence Sinaiticus provides of its setting is that the scribes were familiar with the text of the Greek Bible. Harmonisations to parallel passages and other books take place regularly, showing that the scribes were located in a Christian setting. The main contributions of this study lie in four areas. First, we have increased our knowledge of the actual manuscript and its scribes, filling in some gaps left in the work by Milne and Skeat and simultaneously attempting to open up new areas of research, such as the frequency of the use of nomina sacra and the shape of the Eusebian apparatus. Secondly, our study provides a catalogue of the scribal behaviour of the three people who copied Sinaiticus. It appears that each of them has characteristic habits that work out in a whole range of different areas. In combining textual and codicological studies we have gained a better insight into the work and role of the scribes of Sinaiticus. On the textual level we established the importance of harmonisation to the immediate context and the influence of subsequent text in the creation of copying errors. Likewise we were able to create an overview of other scribal vices. Thirdly, the implications for textual criticism of the Greek Bible are considerable. The testimony of 4 P. McGurk, “The Oldest Manuscripts of the Latin Bible”, in The Early Medieval Bible, ed. R. Gameson, (Cambridge; New York: Cambridge University Press, 1994), 8–10.

Date and Provenance

255

Sinaiticus can now be set against the background of the specific habits of the scribe. For the New Testament, this comes down to a broad confirmation of Westcott and Hort’s description of Sinaiticus as showing “all the ordinary lapses due to rapid and careless transcription” and that Sinaiticus shows a “bold and rough manner of transcription.”5 However, this is characteristic of scribe A and not of scribe D and the sections he wrote. The differences between the scribes are important in the LXX, where we find more text written by scribe D and a number of books written by both scribe D and scribe A. Fourthly, on a methodological level, we have made considerable progress in vindicating the use of singular readings in order to study scribal habits. The method claims to study the work of individual scribes; in the case of Sinaiticus we demonstrated that this is indeed true. The study of Sinaiticus presented here concerned only parts of the manuscript, which leaves areas open for further research. I would like to mention two areas of particular interest. The first is the status of scribe B. We have not studied his singular readings, and neither have we worked with extensive transcripts from his text. Knowing scribe B better could help to resolve the problem of why the text of the prophets in Sinaiticus is not highly esteemed in Septuagint studies. Is this because scribe B is such a poor copyist, or is the text of the exemplar already of inferior quality? A second area for further research is the text of the four gospels. We have seen that the spelling of Matthew does not show scribe A’s normal spelling and something may be going on there that merits investigation, possibly even the identification of another scribal hand.6 An area we did not look into at all is the block-mixture in the first chapters of John, where Sinaiticus has supposedly a Western text. An analysis of scribe A throughout the four Gospels could shed light on the question whether the best explanation for the situation in John is really block-mixture or whether it can be explained by means of the scribal habits of scribe A. How, for example, would scribe A’s text look if he knew the text of John virtually by heart? Besides these questions, much remains to be done in the study of the textual affinities of Westcott-Hort, 246–47. I suggested this in Dirk Jongkind, “‘The Lilies of the Field’ Reconsidered: Codex Sinaiticus and the Gospel of Thomas,” Novum Testamentum 48 (2006): 214, n. 15. 5 6

256

Final Reflections

the correctors. In addition, it will be interesting to see how the new findings from Mt Sinai fit into this current work once they are published. Clearly, there is still plenty left to be studied and discovered and I look forward seeing further studies in Sinaiticus as well as in the other great codices of the Greek Bible.

4 41 -

1

2 3 2 -

4 Macc NS pl. 33 1 3

2 1

28 34 2 1 -

130 14 79 36 40 16 2

10 30 17 1 257

72 4 33 4 24 7

14 10 2 -

Pss Matt Mark NS pl. NS pl. NS pl. 60 1 50 1 44 332 1 54 2 16 16 18 22

2 3 1 26 13

86 13 7 1 38 11 5

90 34 21 21

238 18 2 1 36 2

58 18 14 -

Luke John NS pl. NS pl. 110 80 87 1 45 31 24

APPENDIX I: TABLES AND GRAPHS TO CHAPTER 3

1 104 8

1 Chr NS pl. 53 63 2

Overview of Nomina Sacra of scribe A Scribe A qeos kurios pneuma pna kos ihsous cristos anqrwpo" ourano" uio" dauid ierousalhm

Paul Rev NS pl. NS pl. 628 3 94 434 2 25 164 22 24 216 1 9 389 3 15 122 18 7 - 29 40 10 11 24 2 3 4 2 3 16 1 2 -

Scribe A israhl mhthr pathr swthr

8 1

Appendix I

1 Chr 4 Macc Pss Matt Mark Luke John Paul Rev NS pl. NS pl. NS pl. NS pl. NS pl. NS pl. NS pl. NS pl. NS pl. 46 1 - 29 - 12 2 1 10 4 - 28 2 2 28 1 3 16 7 11 4 1 13 9 1 8 4 7 55 3 11 6 - 42 5 121 17 54 4 2 1 1 1 8 6 -

4

1

11

8 6 16

1

1

2 3

9 6

16 1 1 1

2 1

16 7 14 8

67 45 13 74 3

Appendix I

3

1

2

11 11 13

1

1

2 1 1

1 1 1

2 1 2 1

1

Jdt 4 Macc Pss Matt Mark Luke NS pl. NS pl. NS pl. NS pl. NS pl. NS pl. 70 6 380 11 2 5 10 41 14 1 477 11 8 13 3 1 13 2 1 5 18 12 1 1 3 4 10 1

25

23 1 7 1 2

2 1 1 1

Overview of Nomina Sacra of Scribe D Scribe D qeos kurio" pneuma ihsous cristos anqrwpo" ourano" uio" dauid ierousalhm israhl mhthr pathr swthr staurow

5 1

1 1

Paul NS pl.

1

6 2 3

24 3

2

2

1

1

Rev NS pl. 2 1 3 3

LXX: Itacisms per Folio 120

100

80

60

40

20

0

|

Psalms (scribe D)

Appendix I

|

Other ai-e e-ai ei-i i-ei

Psalms (scribe A)

59 .1 59 .3 59 .5 59 .7 60 .1 60 .3 60 .5 60 .7 61 .1 61 .3 61 .5 61 .7 62 .1 62 .3 62 .5 62 .7 63 .1 63 .3 63 .5 63 .7 64 .1

Synoptic Gospels: Itacisms per folio

XII Prophets

57 .1 57 .3 57 .5 57 .7

120

100

80

60

40

20

0

Appendix I

Other

ai-e

e-ai

ei-i

i-ei

73 .1 73 .3 73 .5 73 .7 74 .1 74 .3 74 .5 74 .7 75 .1 75 .3 75 .5 75 .7 76 .1 76 .3 76 .5 76 .7 77 .1 77 .3 77 .5 77 .7 78 .1 78 .3 78 .5 78 .7

81.7

82.5

Pauline Ep.

81.1

83.3

84.1

84.7

88.3

89.2

90.1

Cath. Ep. Rev

85.5

Appendix I

80.3

Number of paragraphs per folio

79.5

70

78.7

60

78.1

50

75.1

75.7

76.5

4 Gospels

77.3

40

73.5

30

57.3

20

42.7

4 Macc XII Pr.

42.1

10 0 38.7

1 Chr Jdt

34.8

74.3

APPENDIX II: LIST OF THE SUBSTANTIAL DIFFERENCES OF POSITION IN THE EUSEBIAN APPARATUS BETWEEN SINAITICUS AND NA27 Note that in the following table a manuscript can be cited as in agreement with Sinaiticus without agreeing in the table number. In some cases, only the position of the section number of the manuscript agrees. All the data are taken from Swanson’s volumes on the Gospels except those for Sinaiticus, NA27, and the Vulgate.1 In the following tables, the physical location on the folio is indicated by < > (l is line; c is column), and the canonical position by ( ) (v is verse, c is chapter). Other information in the transcript are the position of paragraphs {par}, medial dots {dot}, and in-line spaces {space}. Nomina sacra are indicated by the NS tags [NS] and [/NS]. “Singular” means that Sinaiticus has a unique placement of the apparatus. Matthew Section #NA27 # *ℵ ℵ* Matt 1:18 4/10

Comments

si" *4/10* outw" hn mnhs teuqish" th" mh tro" autou maria" tw iwshf prin h sunelqin autou" eureqh en gastri ecousa ek [NS]pn"[/NS] a giou #4/10# (v 19) iwshf

Singular

1 R.J. Swanson, New Testament Greek Manuscripts (Sheffield: Sheffield Academic Press, 1995); Biblia Sacra iuxta Vulgatam Versionem. 4th ed. (Stuttgart: Deutsche Bibelgesellschaft, 1994).

263

264 Section #NA27 # *ℵ ℵ*

Appendix II

Matt 2:5 (v 5) #5/7# oi de eipan autw 5/7 en bhqleem th" ioudaia" *5/7* outw" gar gegraptai dia tou profhtou Matt 4:3 diabolou (v 2) #16/5# kai nh16/5 steusa" hmera" tesserakonta kai tesserakonta nukta" usteron epinasen {par}(v 3) *16/5* kai proselqwn o pirazwn eipen Matt {par}(v 11) #17/6# tote afihsin au4:11 ton o diabolo" 17/6 *17/6* kai idou aggeloi proshlqon kai Matt kou [NS]isl[/NS] (v 7) #82/2# poreuome10:8 noi de khrussete 82/2 legonte" oti hg giken h basilia twn [NS]ounwn[/NS] (v 8) *82/2* asqe nounta" qerapeu ete leprou" egeireMatt profhtou (v 10) #103/2# outo" 11:10 estin peri ou ge103/2 graptai {par} *103/2* idou egw apostel lw ton aggelon Matt {par}(v 13) *106/10* pante" gar oi pro11:13 fhtai kai o nomo" 106/10 ew" iwannou e profhteusan (v 14) #106/10# kai ei qelete de xasqe auto" estin

Comments Singular, ℵ marked in text

Others: E L f1 2 28(2) 157 700 1346 ℵ marked in text

1 Other: K

All except: NA27; Vulg; CKP

Some others: U 28

All except: NA27; Vulg; D N f1 118

265

Appendix II Section #NA27 # *ℵ ℵ*

Comments

Matt 12:8 116/2

Others: M N Q 124 788

Matt 11:27 112/3

Matt 13:24 136/10

Matt 13:35 140/10

Matt 13:57 142/1

doqh upo tou [NS]pr"[/NS] #112/3# kai oudi" epiginw ski ton [NS]un[/NS] ei mh o [NS]phr[/NS] *112/3* oude ton [NS]pra[/NS] ti" epiginwski ei (v 8) *116/2* [NS]k"[/NS] gar estin tou sab batou o [NS]u"[/NS] tou [NS]anou[/NS] {par}(v 9) #116/2# kai metaba" ekiqen hlqen ei" thn su nagwghn autwn {par}(v 24) #136/10# allhn parabolhn pareqhken autoi" legwn wmoiw qh h basilia twn ouranwn *136/10* [NS]anw[/NS] spiranti kalon sperma en tw a grw autou (v 25) en de tw {par}(v 35) *140/10* opw" plhrwqh to rhqen dia hsa iou tou profhtou legonto" anoi xw en parabole" to stoma mou ereu xomai kekrum mena apo kata bolh" kosmou {par}(v 36) #140/10# tote afi" tou" o clou" eishlqen tauta panta (v 57) #142/1# kai e skandalizonto en autw {par} *142/1* o de eipen autoi" ouk estin profh-

Singular

Singular (1 other) (124) (Not sure of position 124; Swanson clearly wrong on ℵ) Many others except: NA27; Vulg; D L Q f13 1346 1424 124 788

All others (including Vulg) agree with ℵ. NA27 has singular placement

266

Appendix II

Section #NA27 # *ℵ ℵ* Matt 14:24 150/4

Matt 16:2–6 162/5 163/6 164/2

Matt 19:15 193/2

Matt 22:16 223/10 (223/2)

oro" kat idian pro seuxasqe {par} #150/4# oyia" de genome nh" mono" hn e kei (v 24) *150/4* to de ploion hdh meson th" {par}(v 2)#*162/5*# o de apokriqei" ei pen autoi" (v 3)(v 4) #163/6# gene a ponhra kai moi cali" shmion e pizhtei kai shmi on ou doqhsete auth ei mh to sh mion iwna kai ka talipwn autou" aphlqen (v 5) #164/2# kai el qonte" oi maqh tai ei" to peran e pelaqonto artou" labein {par}(v 6) *163/6+2* o de [NS]i"[/NS] eipen orate kai prosecete apo th" zumh" twn farisaiwn kai saddoukaiwn (v 7) #165/6# *164/6* oi de dielogizon [NS]ounwn[/NS] (v 15) *193/2* kai epi qi" ta" cira" ep au tou" eporeuqh e keiqen {par}(v 16) #193/2# kai idou ei" prosel qwn autw eipen {par}(v 15) #223/2# tote poreuqente" oi fariseoi sumbou lion opw" auton pagideuswsin {par}(v 16) *223/10* kai apostellousin autw tou" maqhta"

Comments Others: 118 124 788 579. ℵ marked in text

163/6+2 Near singular: DM *164/6* Others: DLMSYW 1424

Singular ℵ marked in text

Singular

Appendix II Section #NA27 # *ℵ ℵ* Matt 23:32 239/10

Matt 24:16 248/2

Matt 24:30 259/2

Matt 26:11 277/4

Matt 26:21 279/4

Matt 26:45 299/4

tou" profhta" (v 32) *239/10* kai umi" plhrwsate to metron twn pa terwn umwn {par}(v 33) #239/10# ofi" gennhmata ecidnwn skwn noeitw (v 16) *248/2* to te oi en th ioudea feugetwsan epi ta orh (v 17) #248/2# o epi tou dw mato" mh kataba-

sontai (v 30) kai tote fa nhsete to shmion tou [NS]uu[/NS] tou [NS]anou[/NS] en ouranw #259/2# kai koyon te pase e fulai th" gh" *259/2* kai oyonte ton [NS]un[/NS] tou [NS]anou[/NS] erco meq eautwn *277/4* eme de ou pantote e cete (v 12) #277/4# balousa gar auth to muron touto epi tou sw mato" mou pro" (v 20) #279/4# oyia" de genome nh" anekito meta twn ib maqhtwn (v 21) *279/4* kai esqiontwn autwn legi amhn {par}(v 45) #299/4# tote ercete pro" tou" maqhta" kai legi autoi" *299/4* kaqeudete to loipon kai ana-

267 Comments Many others including Vulg. ℵ marked in text All others (including Vulg) agree with ℵ. NA27 has singular placement. ℵ marked in text Most others agree with ℵ; not NA27 Vulg D 157 579 Y U 1424. ℵ marked in text May be just a slight misplacement. Numbers placed before line 24. Singular Others: UP Q 2 1346 124 788

Some others: 788 Q ℵ marked in text

268

Appendix II

Section #NA27 # *ℵ ℵ* Matt 27:2 318/1

Matt 27:11 320/1

Matt 28:4–5 353/2

sai auton (v 2) #318/1# kai dh sante" auton a phgagon *318/1* kai pa redwkan pilatw tw hgemoni (v 11) #320/1# o de [NS]i"[/NS] estaqh em prosqen tou hge mono" *320/1* kai ephrw thsen auton hge mwn legwn su ei kon w" ciwn (v 4) *353/2* apo de tou fobou au tou esisqhsan oi throunte" kai e genhqhsan w" nekroi {par}(v 5) #353/2# apokriqei" de o

Mark Section #NA27 # *ℵ ℵ*

Mark 1:4-10 3/6, 4/1

Mark 1:9 5/1, 4/1

tribou" autou (v 4) #3/6# kai e geneto iwannh" o baptizwn en th ... (v 6) @3/6@ kai hn o iwannh" endedumeno" tri ... agrion (v 7){par}*3/6* #4/1# kai ekhrussen le w (v 9){par} #5/10# *4/1* kai egeneto en ekei nai" tai" hmerai" ... ton iordanhn upo iwannou (v 10) *5/1* kai euqu" anabenwn ek tou

Comments 1 other: Q ℵ marked in text 1 other: 124 ℵ marked in text

Many others agree with ℵ. ℵ marked in text

Comments @3/6@ Y U 700 1424 *3/6*: One other: Q

*4/1*: Singular *5/1*: Others F H Y f1 2 157 700

269

Appendix II Section #NA27 # *ℵ ℵ*

Mark [NS]ounwn[/NS] *6/2* su ei o 1:11, 13 [NS]u"[/NS] 6/2, 7/6 mou o agaphto" en soi eudokhsa (v 12){par} #6/2# kai euqu" to [NS]pna[/NS] au ton ekballei ei" thn erhmon (v 13) kai hn en th erhmw m hmera" pirazo meno" upo tou sa tana *7/6* kai hn meta twn qhriwn kai oi aggeloi dihkonoun autw Mark autw (v 16) kai gramma2:16 tei" twn farisaiwn 23/2 *23/2* kai idonte" oti h sqien meta twn te lwnwn kai amar twlwn elegon toi" maqhtai" autou diati meta twn te lwnwn kai amar twlwn esqiei o di daskalo" umwn {par}(v 17) #23/2# kai akousa" o [NS]i"[/NS] le gei autoi" ou crian Mark {par}(v 23) #33/2# kai proskalesa3:24 meno" autou" en 33/2 parabolai" elegen autoi" pw" duna tai satana" sata nan ekballein (v 24) *33/2* kai ean basileia efeauthn meri" qh ou dunatai sta-

Comments 6/2 Singular 7/6 Many others

Singular

1 other: 579

270

Appendix II

Section #NA27 # *ℵ ℵ*

Comments

Mark 6:6 52/2

2 others: EG ℵ marked in text

Mark 4:2 37/1

Mark 6:15 58/10

Mark 6:21 (6:18) 60/6

*37/1* gh" hsan (v 2) kai edi dasken autou" pol la en parabolai" kai elegen autoi" ... umin to musthri on dedotai th" ba silia" tou [NS]qu[/NS] #37/1# ekei noi" de toi" exw en parabolai" pan ta geinetai (v 12) ina ble sen (v 6) *52/2* kai eqaumasen dia thn apistian autwn #52/2# kai peri hgen o [NS]i"[/NS] ta" kw ma" kuklw dida skwn en autw (v 15) #58/10# alloi de elegon oti hlia" estin {par}*58/10* alloi de oti pro fhth" w" ei" twn profhtwn {par}(v 18) #60/6# elegen gar o iwan nh" tw hrwdh o ti ouk exestin soi ecin thn gunaika tou adelfou sou ... (v 21) *60/6* kai genomenh" h mera" eukairou ote hrwdh" toi" genesioi" autou dipnon epoihsen

Singular, difference one column

2 others: YG (Swanson incorrect on ℵ) 1 other: U

Appendix II Section #NA27 # *ℵ ℵ* Mark 6:34 63/6

Mark 7:6 (7:17) 72/6

Mark 7:26 73/6

{par}(v 32) #63/6# kai aphlqon en ploiw ei" erhmon topon katidian (v 33) kai idon autou" upagonta" kai e pegnwsan autou" polloi {par}kai pezh apo pa swn twn pole wn sunedramon ekei kai prohlqon autou" {par}(v 34) *63/6* kai exelqwn eiden oclon polun kai esplagcnisqh {par}(v 16)(v 17) #72/6# kai ote eishlqon ei" ton oikon apo tou oclou ephrw twn auton oi ma qhtai autou thn parabolhn ... {par}(v 6) *72/6* o de eipen autoi" ka lw" eprofhteusen hsaia" peri umwn twn upokritwn {par}(v 26) *73/6* h de gunh hn ellh ni" surofoinikis sa tw geni #73/6# kai h rwta auton ina to daimonion ek balh ek th" quga tro" auth"

271 Comments 5 others: U U 1424 579 1071

Singular, displacement one column.

Many others. Not: A L U G 788 NA27 Vulg

272

Appendix II

Section #NA27 # *ℵ ℵ* Mark 7:33 (7:36) 75/6 (75/8)

{par}(v 33) *75/6* kai apolabomeno" auton katidian a po tou oclou ela ben tou" daktulou" ... (v 36) kai diestilato au toi" ina mhdeni le gwsin {par} #75/8# oson de autoi" di estelleto autoi mallon perisso terw" ekhrusson Mark (v 37) kai uperperissw" 8:1 exeplhssonto 76/. legonte" #76/6# kalw" panta pepoihken kai tou" kwfou" poiei akouein kai alalou" lalin {par}(c 8)(v 1) *76/.* en ekeinai" tai" h merai" palin pol lou oclou onto" Mark (v 43) #100/6# kai ean skandali9:43 (44 sh se h ceir sou apoor 45) koyon authn kalon 100/6 estin se kullon eiselqein ei" thn zwhn h ta" duo cei ra" econta eisel qein ei" thn geen nan ei" to pur to asbeston {par}(v 44)(v 45) *100/6* kai ean o pou" sou skandalizei se apokoyon auton

Comments Singular

3 (5) others: G 28 1424 (118 579) No canon in ℵ

2 others: 579 (before v 2 44) Q (before v 45)

579 is one section number out in this context and should perhaps be interpreted as a witness to section 101 starting at this point. 2

273

Appendix II Section #NA27 # *ℵ ℵ*

Comments

Mark 10:32 112/2

3 others: M Q 33

Mark 10:17 107/2

Mark 10:46 116/2

Mark 11:27 127/2 Mark 13:3 138/2

Mark 14:7 159/4

ep auta (v 17) #107/2# kai ekpo reuomenou autou *107/2* ei" odon prosdra mwn ei" kai gonu pethsa" auton e {par}(v 32) #112/2# hsan de en th odw anabainonte" ei " ierosoluma kai hn proagwn autou" o [NS]i"[/NS] kai eqamboun to oi de akolouqoun te" efobounto {par} *112/2* kai paralabwn pa lin tou" ib hrxato autoi" legin ta anti pollwn (v 46) #116/2# kai ercontai ei" ieri cw {par} *116/2* kai ekporeuome nou autou apo ie reicw kai twn {par}(v 27) #127/2# kai ercontai palin i" ierosoluma *127/2* kai en tw ierw peri patounto" autou (v 3) #138/2# kai kaqhmenou autou ei" to oro" twn elaiwn katenan ti tou ierou ephrw ta auton {par} *138/2* katidian o petro" kai iakwbo" kai iwannh" kai an sqai eu poihsai *159/4* e me de ou pantote ecete (v 8) #159/4# o escen e poihsen proela ben murisai to sw-

No data; Swanson does not distinguish between the two places

No data; Swanson does not distinguish between the two places Singular

Singular (used in main text as example to illustrate a paragraph attracting the Eus. app.) Singular ℵ marked in text

274

Appendix II

Section #NA27 # *ℵ ℵ*

Comments

Mark 14:32 173/6

Many others

Mark 14:18 161/4

Mark 14:41 180/4

Mark 14:69 196/1

Mark 15:19 208/6

sca (v 17) #161/4#kai oyia" geno menou" ercetai meta twn dwdeka (v 18) *161/4* kai anakeimenwn autwn kai esqion twn o [NS]i"[/NS] eipen te" elegon {dot}(v 32) #*172/1*# kai er contai ei" cwrion ou to onoma geq shmanei #173/6# kai le gei toi" maqhtai" autou *173/6* {dot} kaqisate wde ew" proseu xwmai (v 33) kai para qwsin autw {space}(v 41) #180/4# kai ercetai to triton kai legei autoi" *180/4* kaqeudete to loi pon kai anapau esqe apecei hlqen #196/1# kai exhlqen exw ei" to proaulion (v 69) *196/1* kai h paidiskh idou sa auton hrxato palin legein toi" {par}(v 19) *208/6* kai etupton au tou thn kefalhn kalamw kai ene ptuon autw kai tiqente" ta gona ta prosekunoun autw {par}(v 20) #208/6# kai ote enepaixan autw exedusan au-

Others: G G (C) Q f13 1346

1 other: U Difference is introduction to direct speech 4 others: G 1424 (118) 2

Singular

275

Appendix II Section #NA27 # *ℵ ℵ*

Comments

Luke Section #NA27 # *ℵ ℵ*

Comments

Mark 16:3 231/1

Luke 6:36 (37) 56/2

{par}(v 2) #231/1# kai lian prwi th mia twn sabbatwn ercontai epi to mnhma anatilan to" tou hliou {space} (v 3) *231/1* kai elegon pro" eauta" ti" apoku-

{par}(v 36) #56/2# geinesqe oiktir mone" kaqw" o pathr umwn oi ktirmwn estin (v 37) *56/2* kai mh krinete kai ou mh kriqhte

Luke 7:2 {par}(c 7)(v 1) #65/3# epeidh eplhrwsen 65/3 ta rhmata autou ei" ta" akoa" tou laou eishlqen ei" kafarnaoum{dot} {par}(v 2) *65/3* ekatontarcou de tino" doulo" hmelLuke tou (v 27) #70/2# outo" estin 7:27 peri ou gegraptai 70/2 *70/2* idou apostellw ton aggelon mou Luke {par}(v 18) #81/5(2)# blepetai oun pw" 8:18 akouetai &81/5& o" an 81/5 gar ech doqhse tai autw *81/5* kai o" an mh ech kai o dokei ecin arqhsetai ap autou

Singular

All others (including Vulg) agree with ℵ; NA27 has singular placement. Singular

Others: EHKSQ 565 579 1346 #81/5#: NA27 and many others @81/5@: many others *81/5*: ℵ Singular

276

Appendix II

John Section #NA27 # *ℵ ℵ* John 1:41 17/1

{par}(v 41) #17/1# euriskei outo" prw to" ton adelfon ton idion simwna kai legei autw *17/1* eu rhkamen ton mes sian o estin me qermhneuomeJohn {com par}(v 61) egnw oun [NS]i"[/NS] en 6:61 69/. e autw oti goggu zousin peri tou tou oi maqhtai autou kai eipen autoi" *69/.* touto u ma" skandalizi (v 62) #69/1# ean qewrhtai anabenonta ton [NS]un[/NS] tou anqrwpou opou hn to proJohn (v 41) *83/7* alloi elegon ou7:41 to" estin o [NS]c"[/NS] #83/7# al83/7 loi elegon mh gar ek th" galilaia" o [NS]c"[/NS] ercetai {dot} (v 42) ouci

John 11:54 96/1

*96/1* kakei eminen me ta twn maqhtwn {par}(v 55) #96/1# hn de eggu" to pa sca twn ioudaiwn

Comments 1 other: K

1 other: L ℵ marked in text

Others: L 124 788 1071 1424 E f13 Textual variant: NA27 reads for alloi elegon (2): oi de elegon 1 other: 1071

277

Appendix II Section #NA27 # *ℵ ℵ* John 12:27 107/4

John 13:23 122/1; 123/10

ℵ John 14:15, NA27 John 14:21 129/10

John 14:25 132/10

qitw kai opou ei mi egw ekei kai o diakono" o emo" estai #107/4# ean ti" emoi diakonh seimh sei auton o pathr (v 27) *107/4* nun h yuch mou tetaraktai kai ti sei me (v 22) #122/1# eblepon oun oun oi ioudaioi {correction A, oi ioudaioi: omit} ei" allhlou" oi ma qhtai aporoume noi peri tino" le gei {dot} {par}(v 23) #123//10# *122/10* hn de anakeimeno" {par}(v 15) *129/10* ean agapate ta" en tola" ta" ema" thrhshte (v 16) kagw ... {par}(v 21) o ecwn ta" ento la" mou kai thrwn auta" ekeino" e stin o agapwn me #129/10# o de agapwn me to" me patro" (v 25) *132/10* tau {paragraphus}ta lelalhka umin par umin menwn (v 26) #132/10# o de paraklhto"

Comments All (including Vulg) agree with ℵ except NA27

Singular Section 123 is missing in ℵ.

Others: EGSU Y W 2 28 157 1424 1346 f13

All agree with ℵ except NA27 and Vulg. ℵ marked in text

278

Appendix II

Section #NA27 # *ℵ ℵ*

John 16:14-16 148/3; 149/10

John 18:16 168/1

John 18:28 177/10 John 18:37 181/10

min (v 14) *148/3* ekeino" eme doxasei oti ek tou emou lhmyetai kai anaggelei umin (v 15) #148/3# ... #149/10# ... (v 16) *149/10* meikron kai ouke ti qewreite me kai palin mikron kai

tou arcierew" (v 16) #*166/1*# o de petro" isthkei exw pro" th qura #167/10# exhlqen oun o maqhth" o allo" oshn gnwsto" tw arcierei kai ei pen th qurwrw *168/1* kai eishnegke ton petron (v 17) #168/1# legei oun twrion *177/10* hn de prw i #177/10# kai autoi ouk eishl qon ei" to praitw rion ina mh mian qen {dot}(v 37)*180/4* #180/1# eipen oun autw o pilato" oukoun basileu" ei su {space}*181/10* apekriqh o [NS]i"[/NS] {paragraphus}su legei" oti basi leu" eimi #181/10# egw ei" touto gegennh mai kai ei" touto

Comments *148/3*: 1 other: S *149/10* Others: AGKMSU U P W 2 28 579 1071 148/3 marked in text. Verse 15 missing in ℵ. *168/1*: singular 166/1 present; 167/10 missing in ℵ.

Others: CGHKLU GQ ℵ marked in text. Others: H N f13 157 *181/10* marked in text.

279

Appendix II Section #NA27 # *ℵ ℵ* John 19:2 185/4 John 19:14 194/1

sth" (c 19)(v 1) #185/4# tote oun la bwn o pilato" ton [NS]in[/NS] emastigwsen (v 2) *185/4* kai oi stratiwtai {par}(v 14) *194/1* hn de paraskeuh tou pasca wra hn w" ekth kai legei toi" ioudaioi" i de o basileu" umwn (v 15) #194/1# oi de elegon aron aron staurwson

John 19:25 202/10

klhron #202/10# oi men oun stratiwtai tauta epoihsan (v 25) *202/10* isthki san de para tw stau rw tou [NS]iu[/NS] h mhthr

John 19:31 205/10

(v 31) #205/10# oi oun ioudaioi e pei paraskeuh hn ina mh minh epi tou staurou ta sw mata en tw sabba tw *205/10* hn gar megalh h hmera ekeinou tou sabbatou {dot} hrw-

Comments 1 other: U

1 other: 788 Textual variant: ℵ has shorter text at start of verse 15 All agree with ℵ NA27 has a singular placement. ℵ probably marked in text Singular ℵ either marked with dot or dot was already there.

280

Appendix II

Section #NA27 # *ℵ ℵ* John 21:12 223/9 224/10

John 21:15 227/9; 229/9

Comments

ktuon (v 12) #223/9# legei autoi" o [NS]i"[/NS] deute aristh sate #224/10# *223/9* oudi" de etol ma twn maqhtwn exetasai auton su ti" ei *224/10* eidote" oti o [NS]k"[/NS] estin (v 13) #*225/9*# erce-

*223/9* singular

nekrwn (v 15) ote oun hristhsan legei tw simwni petrw o [NS]i"[/NS] simwn agapa" me pleon toutwn *227/9* legei autw nai [NS]ke[/NS] su oida" oti filw se legei au tw #227/9# boske ta arni a mou {dot}(v 16) #*228/10*# palin le gei autw simwn iwannou agapa" me *229/9* legei autw [NS]ke[/NS] su oida" oti filw se #229/9# legei au tw poimaine ta probata mou {dot}(v 17) #*230/10*# le-

*227/9* Many others

*224/10* singular Discussed in main text. ℵ 225/9 is placed before line 38 and could be taken to indicate a section start at kaiV divdwsin, which is a singular position.

*229/9* Others: L L Y 2 1071 1346 ℵ 229/9 marked in text. Both times the apparatus refers to Peter’s answer instead of Jesus’ question.

APPENDIX III: PLACES WHERE THE EUSEBIAN APPARATUS IS MISSING Section

Place

Comments

67

Mark 6:47

Close to previous number

78

Mark 8:12

81

Mark 8:22

94

Mark 9:33

101

Mark 9:48

113

Mark 10:35

123

Mark 11:19

211

Mark 15:23

213

Mark 15:25

217

Mark 15:29

85

Luke 8:40

107–342

Luke 10:1 – end

31

John 3:36

73

John 6:65

123

John 13:23

167

John 18:16

Except for section 256 at Luke 21:34

Position taken by previous number

281

APPENDIX IV: DEVIATING AND MISSING TABLE NUMBERS IN THE EUSEBIAN APPARATUS Under “Comments” in the table below, “confusion” signifies that the tradition has a variety of options for this table number. “Impossible” means that a table number is mentioned in which the respective Gospel does not figure. The first column gives the Eusebian section number, the second and third column give the table numbers (canons) as found in Sinaiticus and Nestle-Aland 27. Sect.

Table ℵ

Reference

NA 3 Matt 1:1 10 Matt 1:17 5 Matt 3:12

1 2 12

6 5 6

59

6

3

Matt 7:21

77 78 87

5 2 6

6 5 1

Matt 9:36 Matt 9:37 Matt 10:17

131 158

10 2 6 and 5 5

Matt 13:1 Matt 15:24

Comments Singular Singular Others Singular, 1 confusion Near singular Near singular Others Singular 1): Near singular 2): Majority

Apparatus

E 124 788 (E spells out canon 5)

EY UW S P W 28 118 124 788 1) E

The term confusion means that Swanson's data indicate that the manuscript tradition is divergent. 1

283

284

Appendix IV

Sect.

Table

Reference

Comments

Apparatus

163

NA ℵ 5 and 2 6 Matt 16:6

164

6

2

Matt 16:7

194 196 215 216 223

10 2 2 1 10

2 10 6 4 2

Matt 19:21 Matt 19:28 Matt 21:21 Matt 21:22 Matt 22:16

253 323

3 3

2 4

Matt 24:23 Matt 27:16

352

10

1

Matt 28:1

Singular Near singular Singular Near singular Singular, also wrong place Near singular Singular, some confusion Near singular

2

4

1

Mark 1:3

Others

75 76 137 144 188

6 . 6 . 1

8 6 2 2 4

Mark 7:33 Mark 8:1 Mark 13:1 Mark 13:17 Mark 14:54

Singular Missing canon Singular Missing canon Others M G P 28

13

4

1

Luke 3:21

14 68

10 1

3 10

Luke 3:23 Luke 7:17

Singular, impossible Near singular Near singular

74 83

10 3

1 2

Luke 7:36 Luke 8:22

Near singular Singular

14 53

4 1

1 4

John 1:30 John 6:30

Singular Singular?

1) Near singular 2) Near singular Others

1) M 2) E(corrected) E (spelled out) Y L M S W 1424 124 (spelled out) 124 (spelled out)

SW E (spelled out) E S U W 1071 (E spells out canon 2)

A 788 (Swanson wrong on A: lacuna) A L (spelled out)

Swanson defective

285

Appendix IV Sect.

Table ℵ

Reference

NA 10 John 6:38

56

1

69 76 80 88

[.] 10 1 4

1 3 10 1

John 6:61 John 7:28 John 7:33 John 8:20

104 116 118 121 122

3 10 6 1 10

10 3 3 4 1

John 12:24 John 13:13 John 13:16 John 13:21 John 13:23

129

10

1

John 14:15

156 162

4 5

1 1

John 18:1 John 18:12

170

10

1

John 18:20

180

4

1

John 18:37

182 183 187 192 210

. . . 10 .

9 4 4 4 10

John 18:38 John 18:39 John 19:5 John 19:8 John 19:21

201

4

1

John 19:23

Comments

Apparatus

Singular. Section number repeated, should be 57, correct table for 57 Missing canon Near singular N 1071 Near singular L Others N L 124 788 (124 spells also out) Singular Near singular 28 124 788 Near singular 788 Near singular G 124 788 Singular, canon of 123 Others S W 28 (at verse 14: 118) Near singular N Singular, impossible Near singular G N 118 (N also mentioned as 170/4 by Swanson) Others GMNUUP 118 124 788 1424 Missing canon Missing canon Missing canon Near singular N Canon written after section number ( si ) Near singular, 124 788 confusion

286

Appendix IV

Sect.

Table

203

10

219

5

ℵ

Reference

NA 4 John 19:28 9

John 21:1

Comments Near singular, confusion Others, impossible

Apparatus E EL P W

APPENDIX V: UNCORRECTED SINGULAR READINGS IN 1 CHRONICLES Orthography

ℵ

exemplar

1 Chr 9:36, folio 34.8 (OT 3), line 2.10 1 Chr 9:40, folio 34.8 (OT 3), line 2.28 1 Chr 9:40, folio 34.8 (OT 3), line 2.29 1 Chr 9:43, folio 34.8 (OT 3), line 2.43 1 Chr 9:44, folio 34.8 (OT 3), line 2.45 1 Chr 9:44, folio 34.8 (OT 3), line 2.47 1 Chr 9:44, folio 34.8 (OT 3), line 3.5 1 Chr 11:8, folio 34.8 (OT 3), line 6.38 1 Chr 11:11, folio 34.8 (OT 3), line 7.11 1 Chr 11:11, folio 34.8 (OT 3), line 7.12 1 Chr 11:19, folio 34.8 (OT 3), line 8.26 1 Chr 11:22, folio 34.8 (OT 3), line 8.45 1 Chr 11:26, folio 35.1 (OT 4), line 1.36

kir

ki(e)"

iwnaqam

iwnaqan

mareibal

maribaal

rafaian

rafaia

esahl

esehl/ or asahl

ezreikan

esdreikan ?

esahl

asahl ?

oikodomhsen

wkodomhsen

iessaibada

iesebada

acamanni

acamani

pin

piein

bania

banaia

iwa

iwab

287

288

Appendix V

Orthography

ℵ

exemplar

1 Chr 11:26, folio 35.1 (OT 4), line 1.37 1 Chr 11:29, folio 35.1 (OT 4), line 1.43 1 Chr 11:30, folio 35.1 (OT 4), line 1.45 1 Chr 11:31, folio 35.1 (OT 4), line 1.48 1 Chr 11:32, folio 35.1 (OT 4), line 2.5 1 Chr 11:33, folio 35.1 (OT 4), line 2.7 1 Chr 11:35, folio 35.1 (OT 4), line 2.12 1 Chr 11:39, folio 35.1 (OT 4), line 2.21 1 Chr 11:39, folio 35.1 (OT 4), line 2.22 1 Chr 11:40, folio 35.1 (OT 4), line 2.24 1 Chr 11:42, folio 35.1 (OT 4), line 2.28 1 Chr 11:44, folio 35.1 (OT 4), line 2.34 1 Chr 11:46, folio 35.1 (OT 4), line 2.41 1 Chr 11:46, folio 35.1 (OT 4), line 2.43 1 Chr 11:46, folio 35.1 (OT 4), line 2.43 1 Chr 11:47, folio 35.1 (OT 4), line 2.45 1 Chr 11:47, folio 35.1 (OT 4), line 2.45 1 Chr 12:3, folio 35.1 (OT 4), line 3.20

aqlaem

baiqlaem

ia[inc]swqi[/inc]

aswqi

notwfwtei

netwfati ?

rabeiai

rebie ?

garabeqr

garabeqqi

elmaba

elbama

sararei

o arari

ammwneim

ammwni

didwnikeuh

dierwnikeuh ?

ia

ira

eza

saiza

qestarwqei

o astarwqi

aribi

iaribi

el[inc]a[/inc]am

elnaam

kaeqema

kai eqema

iwbhl

iwbhd

eseihl

esseihl

a[inc]i[/inc]abwqei

anaqwqi

289

Appendix V Orthography

ℵ

exemplar

1 Chr 12:5, folio 35.1 (OT 4), line 3.31 1 Chr 12:6, folio 35.1 (OT 4), line 3.32 1 Chr 12:5, folio 35.1 (OT 4), line 3.32 1 Chr 12:6, folio 35.1 (OT 4), line 3.33 1 Chr 12:7, folio 35.1 (OT 4), line 3.39 1 Chr 12:14, folio 35.1 (OT 4), line 4.19 1 Chr 12:28, folio 35.1 (OT 4), line 6.27 1 Chr 12:38, folio 35.1 (OT 4), line 7.46 1 Chr 13:14, folio 35.2 (OT 5), line 2.44 1 Chr 14:4, folio 35.2 (OT 5), line 3.21 1 Chr 14:6, folio 35.2 (OT 5), line 3.25 1 Chr 14:7, folio 35.2 (OT 5), line 3.30 1 Chr 14:11, folio 35.2 (OT 5), line 4.8 1 Chr 14:16, folio 35.2 (OT 5), line 5.7 1 Chr 15:11, folio 35.2 (OT 5), line 6.21 1 Chr 15:18, folio 35.2 (OT 5), line 7.32 1 Chr 15:18, folio 35.2 (OT 5), line 7.34 1 Chr 15:19, folio 35.2 (OT 5), line 7.37

iwzabab

iwzabad

azei

azai

gadara

gadaraqi

ariqmou"

areimouq

soboam

swboam

melcabannea

melcabannaia

twadae

twada"

gaddein

gaddi

abeddaran

abeddaram

samaia

samaa

naget

nageq

enfalet

emfalet

falaad faqeisei

fa(a)l faqiseim

gazaran

gazara

samai anelhm

samaian enhl

makella

makelleia

ieih[inc]a[/inc]

iihl

asab

asaf

290

Appendix V

Orthography

ℵ

exemplar

1 Chr 15:21, folio 35.2 (OT 5), line 7.48 1 Chr 15:24, folio 35.2 (OT 5), line 8.13 1 Chr 15:28, folio 35.3 (OT 6), line 1.7 1 Chr 16:5, folio 35.3 (OT 6), line 2.19 1 Chr 16:16, folio 35.3 (OT 6), line 3.24 1 Chr 16:38, folio 35.3 (OT 6), line 5.34 1 Chr 17:5, folio 35.3 (OT 6), line 7.4 1 Chr 18:3, folio 35.4 (OT 7), line 2.43 1 Chr 19:2, folio 35.4 (OT 7), line 5.30 1 Chr 19:6, folio 35.4 (OT 7), line 6.21 1 Chr 19:7, folio 35.4 (OT 7), line 6.40

enfania"

enfanaia"

sobneia

sobenia

swfeir

swfer

assaf

asaf

isak

isaak

ideqwn

idiqwn

katoikhsa

katwkhsa

adrazarei

adraazar

amman

ammwn

ainan

anan

baidaba

maidaba or badaba

Nonsense words

ℵ

exemplar

1 Chr 12:41, folio 35.1 omoroounte" (OT 4), line 8.19 1 Chr 16:10, folio 35.3 omati (OT 6), line 2.44

omorounte" onomati

Leaps

ℵ

exemplar

1 Chr 9:42, folio 34.8 (OT 3), line 2.39 1 Chr 19:3, folio 35.4 (OT 7), line 5.36

omitted

kai zambri

omitted

enantion sou

291

Appendix V

Add/omit verba minora 1 Chr 10:12, folio 34.8 (OT 3), line 5.4 1 Chr 10:13, folio 34.8 (OT 3), line 5.19 1 Chr 11:10, folio 34.8 (OT 3), line 7.4 1 Chr 11:23, folio 35.1 (OT 4), line 1.8 1 Chr 12:2, folio 35.1 (OT 4), line 3.7 1 Chr 12:19, folio 35.1 (OT 4), line 5.6 1 Chr 12:19, folio 35.1 (OT 4), line 5.12 1 Chr 12:31, folio 35.1 (OT 4), line 6.44 1 Chr 12:33, folio 35.1 (OT 4), line 7.10 1 Chr 12:34, folio 35.1 (OT 4), line 7.18 1 Chr 13:5, folio 35.2 (OT 5), line 1.23 1 Chr 13:8, folio 35.2 (OT 5), line 1.46 1 Chr 16:22, folio 35.3 (OT 6), line 3.48 1 Chr 16:27, folio 35.3 (OT 6), line 4.16 1 Chr 17:26, folio 35.4 (OT 7), line 2.17 1 Chr 19:5, folio 35.4 (OT 7), line 6.13

ℵ

exemplar

swma

to swma

tou kuriou

kuriou

omitted

th

omitted

outo"

en toxw

toxw

omitted

o

omitted

sou

omitted

kai

omitted

autwn

omitted

kai (1)

twn added

oriwn

omitted

kai

toi"

en toi"

to proswpon

proswpon

o auto" qeo"

auto" qeo"

kai ew"

ew"

292

Appendix V

Harmonisation to immediate context 1 Chr 9:35, folio 34.8 (OT 3), line 2.2 1 Chr 10:7, folio 34.8 (OT 3), line 4.11 1 Chr 10:9, folio 34.8 (OT 3), line 4.37 1 Chr 10:14, folio 34.8 (OT 3), line 5.28 1 Chr 11:2, folio 34.8 (OT 3), line 5.38 1 Chr 12:23, folio 35.1 (OT 4), line 6.6 1 Chr 12:32, folio 35.1 (OT 4), line 7.4 1 Chr 12:32, folio 35.1 (OT 4), line 7.6 1 Chr 16:7, folio 35.3 (OT 6), line 2.33 1 Chr 16:15, folio 35.3 (OT 6), line 3.16 1 Chr 16:23, folio 35.3 (OT 6), line 4.5 1 Chr 19:17, folio 35.4 (OT 7), line 8.21

ℵ

exemplar

katwkhsan

katwkhsen

pa" israhl

israhl (2)

autoi"

autwn

kurio"

kurion

basileu"

basilew"

dunamin

dunami"

fulou"

fulh"

ciliade" cilioi kai ciliade" kai oi oi autwn autou mnhmoneuo-mena

mnhmoneuwn

ei" hmera"

ei" hmeran

paretaxanto

paretaxato

Harmonisation to general usage / influence of remote passages 1 Chr 10:2, folio 34.8 (OT 3), line 3.21

ℵ

exemplar reading

melcisedek

melcisoue

Editorial readings

ℵ

exemplar reading

1 Chr 9:33, folio 34.8 (OT 3), line 1.39

efhmeria

efhmeriai

293

Appendix V Editorial readings

ℵ

exemplar reading

1 Chr 9:34, folio 34.8 (OT 3), line 1.47 1 Chr 10:7, folio 34.8 (OT 3), line 4.14 1 Chr 11:2, folio 34.8 (OT 3), line 5.45 1 Chr 14:15, folio 35.2 (OT 5), line 4.46 1 Chr 14:16, folio 35.2 (OT 5), line 5.3 1 Chr 15:23, folio 35.2 (OT 5), line 8.11 1 Chr 19:15, folio 35.4 (OT 7), line 8.1 1 Chr 17:23, folio 35.4 (OT 7), line 2.2

kai arconte"

arconte" (2)

omitted

kai (4)

soi

su

mou

sou

moi

autw

kai pulwroi

pulwroi

autwn

autou

pistwqhto h cir sou

pistwqhto

Nonsense meaning in context 1 Chr 9:40, folio 34.8 (OT 3), line 2.28 1 Chr 10:7, folio 34.8 (OT 3), line 4.19 1 Chr 12:27, folio 35.1 (OT 4), line 6.26 1 Chr 14:10, folio 35.2 (OT 5), line 3.48 1 Chr 15:12, folio 35.2 (OT 5), line 6.30 1 Chr 17:4, folio 35.3 (OT 6), line 7.1 1 Chr 17:5, folio 35.3 (OT 6), line 7.6 1 Chr 17:27, folio 35.4 (OT 7), line 2.25

ℵ

exemplar reading

uiou"

uio"

autoi"

autai"

tetrakiciliai exakosiai anabhtw

tetrakiscilioi exakosioi anabw

htoimasa"

htoimasa

sou

su

anhgagen

anhgagon

soi

su

294

Appendix V

Nonsense meaning ℵ in context 1 Chr 19:13, folio 35.4 poihsai (OT 7), line 7.36

exemplar reading poihsei

Substitutions

ℵ

exemplar reading

1 Chr 10:7, folio 34.8 (OT 3), line 4.18 1 Chr 11:10, folio 34.8 (OT 3), line 7.8 1 Chr 11:23, folio 35.1 (OT 4), line 1.13 1 Chr 12:13, folio 35.1 (OT 4), line 4.16 1 Chr 12:25, folio 35.1 (OT 4), line 6.15 1 Chr 12:25, folio 35.1 (OT 4), line 6.20 1 Chr 12:31, folio 35.1 (OT 4), line 7.1 1 Chr 13:2, folio 35.2 (OT 5), line 1.9 1 Chr 13:7, folio 35.2 (OT 5), line 1.44 1 Chr 13:8, folio 35.2 (OT 5), line 1.47 1 Chr 13:13, folio 35.2 (OT 5), line 2.35 1 Chr 14:10, folio 35.2 (OT 5), line 3.48 1 Chr 14:11, folio 35.2 (OT 5), line 4.12 1 Chr 14:11, folio 35.2 (OT 5), line 4.20 1 Chr 14:14, folio 35.2 (OT 5), line 4.32

katwkisan

katwkhsan

en

epi

wsei

w"

ogdou"

o ogdoo"

uiou

uioi

praxew"

parataxew"

katoikounte"

kat oikou"

uma"

hma"

autwn

autou

enanti

enantion

to

thn kibwton

h

ei

diakoyon

diekoyen

fariein

farasin

poreuqh

poreush

295

Appendix V Substitutions

ℵ

exemplar reading

1 Chr 15:2, folio 35.2 (OT 5), line 5.23 1 Chr 15:24, folio 35.2 (OT 5), line 8.23 1 Chr 15:27, folio 35.2 (OT 5), line 8.40 1 Chr 15:29, folio 35.3 (OT 6), line 1.23 1 Chr 16:6, folio 35.3 (OT 6), line 2.24 1 Chr 16:6, folio 35.3 (OT 6), line 2.26 1 Chr 16:7, folio 35.3 (OT 6), line 2.31 1 Chr 17:16, folio 35.3 (OT 6), line 8.48 1 Chr 17:16, folio 35.4 (OT 7), line 1.2 1 Chr 17:27, folio 35.4 (OT 7), line 2.21 1 Chr 18:16, folio 35.4 (OT 7), line 5.4

alla

all h

th kibwtw

th" kibwtou

perizwsameno"

periezwsmeno"

pezontan

paizonta

enanti

enantion

kuriou

qeou

qeon

kurion

apenantion

apenanti

kurio"

kurie

eulogin

euloghsai

sou" o grammateu"

sousa grammateu"

Transpositions

ℵ

exemplar reading

1 Chr 11:2, folio 34.8 (OT 3), line 5.39 1 Chr 11:13, folio 34.8 (OT 3), line 7.27 1 Chr 12:9, folio 35.1 (OT 4), line 3.47 1 Chr 15:16, folio 35.2 (OT 5), line 7.13 1 Chr 18:7, folio 35.4 (OT 7), line 3.26

eisagagwn ... exagwn ei" polemon ekei

exagwn ... eisagagwn ekei ei" polemon

andr[inc]e"[/inc] dunatoi eufrosunh" en fwnh tou" kloiou" tou" crusou" dauid

dunatoi andre" en fwnh eufrosunh" dauid tou" kloiou" tou" crusou"

296

Add / omit words and clauses 1 Chr 11:4, folio 34.8 (OT 3), line 6.14 1 Chr 12:4, folio 35.1 (OT 4), line 3.23

Appendix V

ℵ

exemplar reading

omitted

ei"

kai berceia kai ihoul o anaqwqei kai samea" o gabawnith" added 1 Chr 12:15, folio 35.1 omitted

(OT 4), line 4.21 1 Chr 12:18, folio 35.1 omitted (OT 4), line 4.46 1 Chr 12:40, folio 35.1 omitted (OT 4), line 8.17

dunato"

gad ei tou paradounai me kai pinonte"

APPENDIX VI: THE CORRECTED READINGS OF PSALMS

Rahlfs / exemplar asebei" epanistantai

oi added

epanistanto

omit oi

Correction

PM?/C?

PM?/C?

Corr. ℵ

epanistantai

Below are listed all the corrections Psalms which are likely made by the earliest scribes, cq. the prima manus. If I had any doubts to whom the actual corrector is, this is indicated in the last column. The second column represents the probable exemplar, the third column the reading originallay found in Sinaitiaticus, and the fourth column the correction. Ps 1:5, folio 59.1 (OT 134), line 1.16 Ps 3:2, folio 59.1 (OT 134), line 2.12

297

298

Ps 15:9, folio 59.3 (OT 136), line 4.46 Ps 27:7, folio 59.7 (OT 140), line 2.39 Ps 29:1, folio 59.7 (OT 140), line 3.34 Ps 48:19, folio 60.5 (OT 146), line 3.47 Ps 60:6, folio 60.8 (OT 149), line 3.28 Ps 67:5, folio 61.1 (OT 150), line 4.48 Ps 67:31, folio 61.2 (OT 151), line 2.32 Ps 68:17, folio 61.2 (OT 151), line 4.11 Ps 68:18, folio 61.2 (OT 151), line 4.17 Ps 71:7, folio 61.3 (OT 152), line 4.2 Ps 73:10, folio 61.4 (OT 153), line 3.17

ℵ

Appendix VI

exomologh-setai

yalmo"

h sarx mou kataskhnwsei ep elpidi hlpisen

[ill]erasure[/ill]

taracqhwsan

exomologhqhsetai o

yalmo" t

hapisen

eisakouson

lawn

taracqhtw-san

to

exomologh-setai

yalmos

PM

PM

PM

PM

PM

PM?/C?

PM

h sarx mou PM kataskhnwsei ep elpidi hlpisen PM

Corr.

kai agalliasqe enwpion autou taracqhson-tai lawn

oisakouson

tacu

PM?/C?

Correction

eisakouson

ta cu+

ou

PM

Rahlfs / exemplar

tacu

tou

o qs+

[ill]erasure[/ill] [ill]erasure[/ill]

ou

omitted

to

o qeo"

Ps 76:5, folio 61.5 (OT 154), line 1.40 Ps 77:4, folio 61.5 (OT 154), line 2.48 Ps 77:8, folio 61.5 (OT 154), line 3.12 Ps 77:49, folio 61.6 (OT 155), line 1.48 Ps 78:1, folio 61.6 (OT 155), line 3.31 Ps 78:6, folio 61.6 (OT 155), line 3.48 Ps 78:9, folio 61.6 (OT 155), line 4.13 Ps 78:11, folio 61.6 (OT 155), line 4.22 Ps 80:6, folio 61.7 (OT 156), line 2.13

epoihsen

pantes oi ecqroi

Correction

PM

PM

PM

Corr.

Appendix VI

o pante" oi ecqroi epoihsan

genwntai

PM

ℵ

epoihsen enwntai

exapesteilen

oi ofqalmoi

genwntai

[ill]erasure[/ill]len

Rahlfs / exemplar

exapesteilen

kata

exelqen

ka

exelqein

kata

sou

restored

klhronomian

PM

PM

PM

PM

PM

klhronomian klhron[ill]erasure [/ill]mian ekceon thn orghn omitted sou epi eqnh ta mh ginwskonta se sou autou

exelqein

299

300 Appendix VI

Corr.

eqeto

PM

apwqeis thn PM a[ill]erasure[/ill]u yuchn mou [ill]erasure[/ill]kateli ouk enkateliyen PM yen PM eba[ill]erasure[/ill]set ebasileusen agalliasetai ai

Correction

eqento

PM?/Ca?

israhl

apwqei" thn yuchn mou ouk egkataleiyei

exhranqhn

PM

isral

ℵ

ebasileusen agalliasqw

exhranqhsan

petrwn

PM

israhl

Rahlfs / exemplar

exhranqhn

pterwn

allhlouia

PM

PM

petrwn

tllhlouia

tw

[ill]erasure[/ill] kai taracqhtwsan ei" ton aiwna [ill]erasure[/ill]wna

Ps 80:14, folio 61.7 (OT 156), line 2.40 Ps 82:18, folio 61.7 (OT 156), line 4.20

allhlouia

katw

PM

PM

tw

arcwn

kai taracqhtwsan eis ton aiwna eqento

Ps 87:7, folio 61.8 (OT 157), line 4.5 Ps 87:15, folio 61.8 (OT 157), line 4.32 Ps 93:14, folio 62.2 (OT 159), line 3.44 Ps 96:1, folio 62.3 (OT 160), line 2.1

arcwn

to logion tou kuriou epurwsen auton added

Ps 101:12, folio 62.4 (OT 161), line 1.38 Ps 103:12, folio 62.5 (OT 162), line 1.2 Ps 104:1, folio 62.5 (OT 162), line 2.23 Ps 104:1, folio 62.5 (OT 162), line 2.24 Ps 104:20, folio 62.5 (OT 162), line 3.30

Ps 105:7, folio 62.6 (OT 163), line 1.34 Ps 105:10, folio 62.6 (OT 163), line 1.44 Ps 108:4, folio 62.7 (OT 164), line 3.48 Ps 108:12, folio 62.7 (OT 164), line 4.26 Ps 111:8, folio 62.8 (OT 165), line 3.33 misountwn

parepikranan

foboum added

endiaballon

[ill]erasure[/ill]twn

parapikranan

thn

ou fobhqhsetai

orfanois

endieballon

misountwn

parepikranan

Correction

PM

PM

PM

PM

PM

PM

PM

Corr.

Appendix VI

endieballon

ou fobhqhsetai etoimh h kardia autou elpizein epi ton kurion tahn

to

PM

ℵ

orfanoi"

ei" added

omitted

Rahlfs / exemplar

ou mh fobhqh

sou

Ps 118:36, folio 63.2 thn (OT 167), line 2.2 Ps 118:38, folio 63.2 to (OT 167), line 2.7 Ps 118:69, folio 63.2 mou (OT 167), line 3.36

301

302

Ps 118:70, folio 63.2 (OT 167), line 3.38 Ps 118:82, folio 63.2 (OT 167), line 4.24 Ps 118:88, folio 63.2 (OT 167), line 4.37 Ps 118:92, folio 63.2 (OT 167), line 4.45 Ps 118:127, folio 63.3 (OT 168), line 2.38 Ps 118:134, folio 63.3 (OT 168), line 3.8 Ps 118:141, folio 63.3 (OT 168), line 3.25 Ps 118:149, folio 63.3 (OT 168), line 3.47 Ps 118:163, folio 63.3 (OT 168), line 4.33 ei

pote parakalesei" tou

to

i

sou

parekalesan

autouwn

fulaxw

omitted

ei

pote paralaleseis tou

autwn

Correction

PM

PM

PM

PM

PM

PM

Corr.

Appendix VI

sou

fulaxai

newteros

PM

ℵ

fulaxw

newtero

eleos

PM

autwn

newtero"

logion

kai

Rahlfs / exemplar

eleo"

omitted

PM

kai

Ps 120:8, folio 63.4 (OT 169), line 2.3 Ps 124:4, folio 63.4 (OT 169), line 3.30 Ps 127:3, folio 63.4 (OT 169), line 4.33 Ps 130:1, folio 63.5 (OT 170), line 1.39 Ps 130:2, folio 63.5 (OT 170), line 1.41 Ps 131:5, folio 63.5 (OT 170), line 2.10 Ps 131:13, folio 63.5 (OT 170), line 2.35 Ps 133:1, folio 63.5 (OT 170), line 3.17 Ps 134:6, folio 63.5 (OT 170), line 3.34 nqofuta

euqe" added

neofuta

omitted

fulaxei

Correction

PM

PM

PM

PM

PM+Ca

Corr.

Appendix VI

agaqoi"

omitted

restores original line order restores original line order k_w

PM

oude eporeuqhn en megaloi" alla

fulaxei"

neofuta

oude eporeuqhn en megaloi" kai w

elelexato

PM

ℵ

kuriw

exelaxato

en oikw kuriou

PM

fulaxei

exelexato

omitted

qalassais

Rahlfs / exemplar

en oikw kuriou

alassai"

PM

qalassai"

303

304

Ps 134:8, folio 63.5 (OT 170), line 3.38 Ps 134:10, folio 63.5 (OT 170), line 3.43 Ps 138:6, folio 63.6 (OT 171), line 3.11 Ps 139:2, folio 63.6 (OT 171), line 4.20 Ps 142:2, folio 63.7 (OT 172), line 3.2 Ps 144:13, folio 63.8 (OT 173), line 1.40 Ps 147:1, folio 63.8 (OT 173), line 3.41 Ps 148:1, folio 63.8 (OT 173), line 4.17 authn

o" epataxen

andikou

auton

apektinen

tou doulou

adikou

authn

epataxen

prototoka

Correction

PM

PM

PM

PM

PM

PM

Corr.

Appendix VI

adikou

twn doulwn

basileia

PM?/C?

prwtotoa

tou doulou

baleia

ainei

PM

ℵ

basileia

ainei [inc]q[/inc]

ton

prwtotoka

ainei

tan

Rahlfs / exemplar

ton

BIBLIOGRAPHY Libri Salomonis id est Proverbia Ecclesiastes Canticum Canticorum. Vol. XI Biblia Sacra iuxta Latinam vulgatam versionem ad codicum fidem iussu Pii PP. XII, cura et studio monachorum abbatiae pontificiae Sancti Hieronymi in urbe ordinis sancti Benedicti edita. Rome: Typis polyglottis Vaticanis, 1957. Abbot, Ezra. "On the Comparative Antiquity of the Sinaitic and Vatican Manuscripts". Journal of the American Oriental Society 10 (1880): 189–200. Aland, Barbara. "Das Zeugnis der frühen Papyri für den Text der Evangelien: diskutiert am Matthäusevangelium". Pages 1, 325–35 in The Four Gospels 1992. Edited by F. Van Segbroeck, et al. Leuven: University Press; Peeters, 1992. — "Kriterien zur Beurteilung kleinerer Papyrusfragmente des Neuen Testaments". Pages 1–13 in New Testament Textual Criticism and Exegesis. Edited by A. Denaux. Leuven: University Press; Peeters, 2002. Aland, Kurt. Kurzgefasste Liste der Griechischen Handschriften des Neuen Testaments. 2 ed. Arbeiten zur Neutestamentlichen Textforschung 1. Berlin; New York: Walter De Gruyter, 1994. — and Barbara Aland. The Text of the New Testament: An Introduction to the Critical Editions and to the Theory and Practice of Modern Textual Criticism. 2nd ed. Grand Rapids, MI; Leiden: W.B. Eerdmans; E.J. Brill, 1989. American and British committees of the International Greek New Testament Project. The Gospel according to St. Luke. Vol. 1, Chapters 1– 12. The New Testament in Greek 3. Oxford: Clarendon, 1984. Balogh, J. "Voces Paginarum". Philologus 82 (1927): 84–109, 202–40. Bate, H.N. "Luke xxii 40". Journal of Theological Studies 36 (1935): 76–77. Beneshevich, V.N. Catalogus Codicum Manuscriptorum Graecorum qui in Monasterio Sanctae Catherinae in Monte Sina Asservantur. St Petersburg: 1911. Repr. Hildesheim: Olms, 1965. — . Les manuscrits Grecs du Mont Sinaï et le monde savant de l'Europe depuis le XVIIe siècle jusqu'à 1927. Texte und Forschungen zur ByzantinischNeugriechischen Philologie; Beiheft 21. Athens: Verlag der "Byzantinisch-Neugriechischen Jahrbücher", 1937. 305

306

Appendix VI

Berger, Samuel. Histoire de la Vulgate pendant les premiers siècles du moyen âge. Paris: Hachette et cie, 1893. Blass, F., A. Debrunner, and Robert F. Funk. A Greek Grammar of the New Testament and Other Early Christian Literature. Cambridge, Chicago, Ill.: Cambridge University Press, University of Chicago Press, 1961. Böttrich, C. Bibliographie Konstantin von Tischendorf (1815–1874). Leipzig: Leipziger Universitätsverlag, 1999. Brooke, A.E., Norman McLean, and Henry St John Thackeray. The Old Testament in Greek: according to the Text of Codex Vaticanus, Supplemented from other Uncial Manuscripts, with a Critical Apparatus Containing the Variants of the Chief Ancient Authorities for the Text of the Septuagint. Vol. II: The Later Historical Books, part III: I and II Chronicles. Cambridge: Cambridge University Press, 1932. Bruin, W.M. de. "Interpreting Delimiters: The Complexity of Text Delimitation in Four Major Septuagint Manuscripts". Pages 66–89 in Studies in Scriptural Unit Division. Edited by M.C.A. Korpel and J.M. Oesch. Assen: Van Gorcum, 2002. Burgon, John W. The Last Twelve Verses of the Gospel According to S. Mark Vindicated against Recent Critical Objectors. Oxford; London: James Parker and Co, 1871. Canart, Paul. "Notice Paléographique et codicologique". Pages 1–6 in Prolegomena to Bibliorum Sacrorum Graecorum Codex Vaticanus B. Roma: Istituto Poligrafico e Zecca dello Stato, 1999. Caragounis, Chrys C. The Development of Greek and the New Testament: Morphology, Syntax, Phonology, and Textual Transmission. Wissenschaftliche Untersuchungen zum Neuen Testament 167. Tübingen: Mohr Siebeck, 2004. Cavallo, Guglielmo. Ricerche sulla maiuscola biblica. Firenze: Le Monnier, 1967. Colwell, Ernest C. "Significance of Grouping of New Testament Manuscripts". New Testament Studies 4 (1958): 73–92. — . "Scribal Habits in Early Papyri: A Study in the Corruption of the Text". Pages 370–89 in The Bible in Modern Scholarship. Nashville: Abingdon Press, 1965. — . "Hort Redivivus: A Plea and a Program". Pages 131–55 in Transitions In Biblical Scholarship. Chicago: University of Chicago Press, 1968. — . Studies in Methodology in Textual Criticism of the New Testament. New Testament Tools and Studies 9. Grand Rapids: Eerdmans, 1969. Comfort, P. "Scribes as Readers: Looking at New Testament Textual Variants According to Reader Reception Analysis". Neotestamentica 38 (2004): 28–53.

Appendix VI

307

Crowder, Robert G., and Richard K. Wagner. The Psychology of Reading: An Introduction. New York; Oxford: Oxford University Press, 1992. Dain, Alphonse. Les manuscrits. Collection d'études anciennes. 3rd edition. Paris: Les Belles Lettres, 1975. Epp, Eldon Jay, and Gordon D. Fee. Studies in the Theory and Method of New Testament Textual Criticism. Studies and Documents 45. Grand Rapids, Mich.: W.B. Eerdmans, 1993. Fee, Gordon D. "Codex Sinaiticus in the Gospel of John: a Contribution to Methodology in Establishing Textual Relationships". In Studies in the Theory and Method of New Testament Textual Criticism, ed. Eldon Jay Epp and Gordon D. Fee, 221–43. Grand Rapids, Mi.: Eerdmans, 1993. Fischer, Bonifatius, Jean Gribomont, H.F.D. Sparks, Walter Thiele, Robert Weber, Hermann Josef Frede, and Roger Gryson. Biblia Sacra iuxta Vulgatam Versionem. 4th ed. Stuttgart: Deutsche Bibelgesellschaft, 1994. Fraenkel, Detlef. Verzeichnis der griechischen Handschriften des Alten Testaments. Septuaginta: Vetus Testamentum Graecum, Supplementum vol. I, 1: Die Überlieferung bis zum VIII. Jahrhundert. Göttingen: Vandenhoeck & Ruprecht, 2004. Gamble, Harry Y. Books and Readers in the Early Church: A History of Early Christian Texts. New Haven; London: Yale University Press, 1995. Gardthausen, Viktor Emil. Griechische Palaeographie. 2 vols. Leipzig: Veit & Comp., 1911–13. Gignac, Francis Thomas. A Grammar of the Greek Papyri of the Roman and Byzantine Periods. Vol. 1 - Phonology. Testi e documenti per lo studio dell'antichita 55. Milano: Istituto Editoriale Cisalpino - La Gollardica, 1976. — . A Grammar of the Greek Papyri of the Roman and Byzantine Periods. Vol. 2 Morphology. Testi e documenti per lo studio dell'antichità 55. Milano: Istituto editoriale cisalpino - La Goliardica, 1981. Gregory, Caspar Rene. "The Quires in Greek Manuscripts". American Journal of Philology 7 (1886): 27–32. — . Textkritik des Neuen Testamentes. Vol. 1. Leipzig: J.C. Hinrichs, 1900. — . Canon and Text of the New Testament. International Theological Library. Edinburgh: T. & T. Clark, 1907. Hanhart, R. Text und Textgeschichte des 2. Esrabuches. Abhandlungen der Akademie der Wissenschaften zu Göttingen, Philologisch-Historische Klasse; 3. Folge, Bd. 253. Göttingen: Vandenhoeck & Ruprecht, 2003. Harris, J. Rendel. "An Important Ms. of the New Testament (cod. Mus. Brit. Add. 33277)". Journal of Biblical Literature 9 (1890): 31–59. — . Stichometry. London: C.J. Clay and sons, 1893.

308

Appendix VI

Head, Peter M. "Observations on Early Papyri of the Synoptic Gospels, Especially on the 'Scribal Habits'". Biblica 71 (1990): 240–247. — . "The Habits of New Testament Copyists: Singular Readings in the Early Fragmentary Papyri of John". Biblica 85 (2004): 399–408. Hengel, Martin, and Roland Deines. The Septuagint as Christian Scripture: Its Prehistory and the Problem of its Canon. Old Testament Studies. Edinburgh; New York: T & T Clark, 2002. Hurtado, Larry H. "The Earliest Evidence of an Emerging Christian Material and Visual Culture: the Codex, the Nomina Sacra and the Staurogram". Pages 271–88 in Text and Artifact in the Religions of Mediterranean Antiquity: Essays in Hounour of Peter Richardson. Edited by Stephen G. Wilson and Michel Desjardins. Canadian Corporations for Studies in Religion, 2000. — . "The Origin of the Nomina Sacra: a Proposal". Journal of Biblical Literature 117 (1998): 655–73. Iera Monh kai Archiepiskope Sina. Ta Nea Euramata. Athens: Idruma orous Sina, 1998. Irigoin, Jean. "Les cahiers des manuscrits grecs". Pages 1–19 in Recherches de codicologie comparée: la composition du codex au Moyen Âge, en Orient et en Occident. Edited by Philippe Hoffmann. Paris: Pr. de l'École normale supérieure, 1998. Jongkind, Dirk. "Dissertation Summary: Studies in the Scribal Habits of Codex Sinaiticus". Tyndale Bulletin 56 (2005): 153–56. — . "One Codex, Three Scribes, and Many Books: Struggles with Space in Codex Sinaiticus". Pages 121–35 in New Testament Manuscripts: Their Texts and Their World, ed. Thomas J. Kraus and Tobias Nicklas. Texts and Editions for New Testament Study 2. Leiden; Boston: Brill, 2006. — ."'The Lilies of the Field' Reconsidered: Codex Sinaiticus and the Gospel of Thomas." Novum Testamentum 48 (2006): 209–16. Junack, Klaus. "Abschreibpraktiken und Schreibergewohnheiten in ihrer Auswirkung auf die Textüberlieferung". Pages 277–95 in New Testament Textual Criticism: Its Significance for Exegesis. Essays in honour of Bruce M. Metzger. Edited by Eldon Jay Epp and Gordon D. Fee. Oxford: Clarendon Press, 1981. Kasser, Rodolphe. Papyrus Bodmer XXIV: Psaumes XVII-CXVIII. ColognyGenève: Bibliotheca Bodmeriana, 1967. Knox, Bernard M.W. "Silent Reading in Antiquity". Greek, Roman, and Byzantine Studies 9 (1968): 421–35.

Appendix VI

309

Korpel, Marjo Christina Annette, and Joseph M. Oesch, eds. Delimitation Criticism: A New Tool in Biblical Scholarship. Pericope 1. Assen: Van Gorcum, 2000. — . Studies in Scriptural Unit Division. Pericope 3. Assen: Van Gorcum, 2002. Kraft, R.A. "The 'Textual Mechanics' of Early Jewish LXX/OG Papyri and Fragments". Pages 51–72 in The Bible as Book: The Transmission of the Greek Text. Edited by Scot McKendrick and Orlaith A. O'Sullivan. London: The British Library, 2003. Lake, Helen, and Kirsopp Lake. Codex Sinaiticus Petropolitanus: the New Testament, the Epistle of Barnabas and the Shepherd of Hermas,Preserved in the Imperial Library of St. Petersburg. Oxford: Clarendon Press, 1911. Lake, Kirsopp, and Helen Lake. Codex Sinaiticus Petropolitans et FridericoAugustanus Lipsiensis: the Old Testament Preserved in the Public Library of Petrograd, in the Library of the Society of Ancient Literature in Petrograd, and in the Library of the University of Leipzig. Oxford: Clarendon Press, 1922. Lake, Kirsopp. "Review of 'Scribes and Correctors of the Codex Sinaiticus' ". Classical Philology 37 (1942): 91–96. McGurk, Patrick. Latin Gospel books from A.D. 400 to A.D. 800. Les publications de scriptorium 5. Paris: Editions "Erasme" StandaardBoekhandel, 1961. — . "The Oldest Manuscripts of the Latin Bible". Pages 1–23 in The Early Medieval Bible: Its Production, Decoration and Use. Edited by Richard Gameson. Cambridge; New York: Cambridge University Press, 1994. Metzger, Bruce M. Manuscripts of the Greek Bible: An Introduction to Palaeography. New York: Oxford University Press, 1981. — . The Text of the New Testament: Its Transmission, Corruption and Restoration. 3rd ed. New York; Oxford: Oxford University Press, 1992. Milne, Herbert John Mansfield, T.C. Skeat, with contributions by Douglas Cockerell. Scribes and Correctors of the Codex Sinaiticus. London: British Museum, 1938. — . Notes on Codex Sinaiticus. Manuscript held in the British Library, Add. Ms. 78935 A. Signed off by the authors December 14, 1939. Mondrain, Brigitte. "Les signatures des cahiers dans les manuscrits grecs". Pages 21–48 in Recherches de codicologie comparée: la composition du codex au Moyen Âge, en Orient et en Occident. Edited by Philippe Hoffmann. Paris: Pr. de l'École normale supérieure, 1998. Morton, A.Q. "Codex Sinaiticus Revisited". Irish Biblical Studies 24 (2002): 14–31. Nestle, E. "Die Eusebianische Evangelien-Synopse". Neue kirchliche Zeitschrift 19 (1908): 40–51, 93–114, 219–32.

310

Appendix VI

Nordenfalk, Carl. Die spätantiken Kanontafeln: kunstgeschichtliche Studien über die eusebianische Evangelien-konkordanz in den vier ersten Jahrhunderten ihrer Geschechte. Goteborg: O. Isacsons boktryckeri a.-b., 1938. — . "The Eusebian Canon-tables: Some Textual Problems". Journal of Theological Studies 35 (1984): 96–104. O'Callaghan, Jose. Nomina sacra in papyris Graecis saeculi III neotestamentariis. Analecta biblica 46. Rome: Biblical Institute Press, 1970. — . "'Nominum Sacrorum' elenchus in Graecis Novi Testamenti papyris a saeculo IV usque ad VIII". Studia Papyrologica 10 (1971): 99–122. Paap, A.H.R.E. Nomina Sacra in the Greek Papyri of the First Five Centuries AD: The Sources and Some Deductions. Papyrologica Lugduno-Batava 8, 1959. Parker, D.C. Codex Bezae: An Early Christian Manuscript and its Text. Cambridge: CUP, 1992. — . "The Majuscule Manuscripts of the New Testament". Pages 22–42 in The Text of the New Testament in Contemporary Research: Essays on the Status Quaestionis. Edited by Bart D. Ehrman and Michael W. Holmes. Grand Rapids, Mi.: Eerdmans, 1995. Parry, D.W., D.V. Arnold, D.G. Long, and S.R. Woodward. "New Technological Advances: DNA, Databases, Imaging Radar". Pages I.496–515 in The Dead Sea Scrolls after Fifty Years. A Comprehensive Assessment. Edited by P.W. Flint and J.C. VanderKam. Leiden: Brill, 1998. Pattie, T.S. "The Codex Sinaiticus". The British Library Journal 3 (1977): 1–6. — . "The Creation of the Great Codices". Pages 61–72 in The Bible as Book: The Manuscript Tradition. Edited by John L. Sharpe III and Kimberly Van Kampen. London; New Castle: British Library; Oak Knoll Press, 1998. P. Petitmengin and B. Flusin, "Le Livre Antique et la Dictée: Nouvelles recherches". Pages 247–62 in Mémorial André-Jean Festugière: Antiquité paienne et chrétienne. Edited by E. Lucchesi and H.D. Saffrey. Genève, 1984. Pietersma, Albert. Two Manuscripts of the Greek Psalter in the Chester Beatty Library, Dublin. Analecta biblica 77. Rome: Biblical Institute Press, 1978. — . "The Edited Text of P.Bodmer XXIV". Bulletin of the American Society of Papyrologists 17 (1980): 67–79. Porfiri, P. Vostok khristianskii: Egipet i Sinai; bidy, ocherki, plany i nadpisi. 2 vols. St Petersburg, 1857.

Appendix VI

311

Rahlfs, A. "Alter und Heimat der vaticanischen Bibelhandschrift". Nachrichten von der Königl. Gesellschaft der Wissenschften zu Göttingen. Philologisch-historische Klasse (1899): 72–79. — . "Curiosa im Codex Sinaiticus". Zeitschrift für die alttestamentliche Wissenschaft 50 (1932): 309–10. Royse, J.R. "Scribal Habits in Early Greek New Testament Papyri". unpublished doctoral dissertation, Graduate Theological Union, 1981. Rudberg, Gunnar. Neutestamentlicher Text und Nomina Sacra. Skrifter utgifna af K. Humanistiska Vetenskaps-Samfundet i Upsala. 17:3. Uppsala, Leipzig: A.-B. Akademiska Bokhandeln, Harrassowitz, 1915. Sanders, H.R. "Review of 'Scribes and Correctors of Codex Sinaiticus', H.J.M Milne and T.C. Skeat". American Journal of Philology 60 (1939): 486–90. Scrivener, F.H.A. A Full Collation of the Codex Sinaiticus with the Received Text of the New Testament. Cambridge, 1864. — , and Edward Miller. A Plain Introduction to the Criticism of the New Testament for the Use of Biblical Students. London: G. Bell, 1894. Ševčenko, I. "New Documents on Tischendorf and the Codex Sinaiticus". Scriptorium 18 (1964): 55–80. Skeat, T.C. The Use of Dictation in Ancient Book Production. Proceedings of the British Academy 1956, 1957. — . "The 'Codex Sinaiticus', the 'Codex Vaticanus' and Constantine". Journal of Theological Studies 50 (1999): 583–625. — . The Collected Biblical Writings of T.C. Skeat, Introduced and Edited by J.K. Elliott. Supplements to Novum Testamentum CXIII. Leiden, Boston: Brill, 2004. Soden, Hermann Freiherr von. Die Schriften des neuen Testaments: in ihrer ältesten erreichbaren Textgestalt I. Teil: Unterschungen; I. Abteilung: Die Textzeugen. Göttingen: Vandenhoeck und Ruprecht, 1911. Stunt, T.C.F. "Some Unpublished Letters of S P Tregelles Relating to the Codex Sinaiticus". Evangelical Quarterly 48 (1976): 15–26. Swanson, Reuben J. New Testament Greek Manuscripts: Variant Readings Arranged in Horizontal Lines against Codex Vaticanus. 4 vols.: Matthew, Mark, Luke, John. Sheffield: Sheffield Academic Press, 1995. Thackeray, Henry St John. A Grammar of the Old Testament in Greek. Cambridge: University Press, 1909. Thompson, Sir Edward Maunde. An Introduction to Greek and Latin Palaeography. Oxford: Clarendon Press, 1912. Tindall, Christian. Contributions to the Statistical Study of the Codex Sinaiticus. Edinburgh: Oliver and Boyd, 1961.

312

Appendix VI

Tischendorf, A.F.C. Codex Friderico-Augustanus: sive fragmenta Veteris Testament e codice Graeco omnium qui in Europa supersunt facile antiquissimo in oriente detexit in patriam attulit ad modum codicis edidit C. Tischendorf. Leipzig, 1846. — . Fragmenta sacra palimpsesta sive fragmenta cum Novi tum Veteris Testamenti ex quinque codicibus graecis palimpsestis antiquissimis nuperrime in oriente repertis. Addita sunt fragmenta psalmorum papyracea et fragmenta evangelistariorum palimpsesta, item fragmentum codicis Friderico-Augustani. Nunc primum eruit atque edidit Aenoth. Frideric. Constantinus Tischendorf. Monumenta sacra inedita. Nova collectio 1. Leipzig: J.C. Hinrichs, 1855. — . Bibliorum codex Sinaiticus Petropolitanus: Ex tenebris protraxit in Europam transtulit ad juvandas atque illustrandas sacras litteras edidit C. Tischendorf. 4 vols. St Petersburg, 1862. — . Novum Testamentum Sinaiticum: sive Novum Testamentum cum epistula Barnabae et fragmentis Pastoris ex codice Sinaitico. Leipzig: F.A. Brockhaus, 1863. — . Novum Testamentum Graece: Ex Sinaitico codice omnium antiquissimo Vaticana itemque Elzeviriana lectione notata edidit A.F.C. Tischendorf. Leipzig, 1865 (Reprinted Olms: Hildesheim, 1969). — . When Were our Gospels Written? With a Narrative of the Discovery of the Sinaitic Manuscript. London: Religious Tract Society, 1896. Tov, E. Scribal Practices and Approaches Reflected in the Texts Found in the Judean Desert. Studies on the Texts of the Desert of Judah 54, ed. F. Garcia Martinez. Leiden; Boston: Brill, 2004. Traube, Ludwig. Nomina Sacra: Versuch einer Geschichte der christlichen Kurzung. Quellen und Untersuchungen zur lateinischen Philologie des Mittelalters Bd. 2. München: C.H. Beck'sche Verlag., 1907. Turner, E.G. The Typology of the Early Codex. Haney Foundation Series 18. University of Pennsylvania [Philadelphia, Pa.]: University of Pennsylvania Press, 1977. Underwood, G., and V. Batt. Reading and Understanding: An Introduction to the Psychology of Reading. Oxford: Blackwell, 1996. Westcott, B.F., and F.J.A. Hort. The New Testament in the Original Greek: Introduction, Appendix. 2nd ed. London: Macmillan, 1896. Wordsworth, J., H.J. White, H.F.D. Sparks, and A.W. Adams. Nouum Testamentum Domini Nostri Iesu Christi Latine secundum editionem Sancti Hieronymi. Part Three: Acts - Catholic Epistles - Apocalyps of John. Oxonii: E Typographeo Clarendoniano, 1954.

INDEX OF MODERN AUTHORS No attempt has been made to provide references to Milne-Skeat or to any of T.C. Skeat’s other contributions; they are present on almost any page of this study. Abbot, E., 18 Aland, B., 134, 140 Balogh, J., 23 Bate, H.N., 24 Beneshevich, V.N., 6, 7, 51 Berger, S., 121 Burgon, J.W., 18 Canart, P., 31, 35, 41 Caragounis, C.C., 24, 91 Cavallo, G., 1, 13, 20 Colwell, E.C., 135, 140, 242 Comfort, P., 140 Dain, A., 23, 136, 143, 251 De Bruin, W., 100 Fee, G.D., 25 Flusin, B., 21 Gamble, H.Y., 24 Gardthausen, V.E., 18 Gregory, C.R., 31 Harris, J.R., 19, 222 Head, P.M., 140 Hort, F.J.A.. See Westcott-Hort Hurtado, L.W., 62, 67 Irigoin, J., 31 Junack, K., 23, 136, 251 Knox, B.M.W., 23 Kraft, R.A., 95 Lake, Kirsopp, 2, 7, 11, 12, 13, 16, 17, 18, 19, 20, 21, 22, 47 McGurk, P., 31, 33, 254

Metzger, B.M., 13, 23, 84, 96, 139, 140 Mondrain, B., 32 Morton, A.Q., 26 Nestle, E., 110, 118, 119 Nordenfalk, C., 110, 111 Parker, D.C., 18, 19, 36, 63 Pattie, T.S., 23, 34 Petitmengin, P., 21 Porfiri Uspenski, 6 Rahlfs, A., 19, 21, 25 Royse, J.R., 134, 136, 137, 138, 139, 140, 141, 144, 246 Sanders, H.A., 20, 22, 95 Scrivener, F.H.A., 7, 9, 10, 11, 12, 32, 41, 90, 132 Stunt, T.C.F., 6 Thompson, E.M., 18, 84 Tindall, C., 25, 46 Tischendorf, A.F.C., 1, 2, 5, 6, 7, 9, 10, 11, 12, 16, 17, 18, 24, 33, 47, 48, 63, 64, 65, 84, 102, 104, 131, 132, 135, 146, 216, 222, 230 Tov, E., 95 Traube, L., 62, 63, 83, 84 Tregelles, S.P., 6, 10 von Soden, H., 24, 122 Westcott-Hort, 24, 105, 121, 131, 132, 133, 134, 255

313

INDEX OF SCRIPTURAL PASSAGES Note that all references for Psalms are to the Septuagintal verse numbering. 1 Chr 9:27–19:17, 145 9:29, 157, 161 9:32, 157 9:33, 150, 292 9:34, 293 9:35, 292 9:36, 287 9:37, 147 9:38, 155, 161 9:40, 287, 293 9:42, 290 9:43, 287 9:44, 157, 287 10:2, 154, 292 10:7, 153, 154, 155, 292, 293, 294 10:8, 152 10:9, 292 10:10, 157 10:12, 151, 160, 291 10:13, 156, 163, 291 10:14, 292 11:2, 150, 156, 158, 161, 163, 292, 293, 295 11:4, 151, 152, 296 11:6, 151, 160

11:8, 287

11:9, 153 11:10, 291, 294 11:11, 153, 287 11:13, 295 11:18, 158 11:19, 154, 161, 287 11:20, 153 11:22, 8, 287 11:23, 150, 291, 294 11:26, 287, 288 11:29, 147, 288 11:30, 288 11:31, 147, 288 11:32, 288 11:33, 148, 288 11:35, 288 11:39, 288 11:40, 288 11:41, 156 11:42, 288 11:43, 148 11:44, 148, 288 11:46, 288 11:47, 288 12:2, 156, 291 12:3, 288 12:4, 148, 158, 296 314

12:5, 148, 289 12:6, 289 12:7, 289 12:9, 295 12:13, 294 12:14, 289 12:15, 296 12:16, 158 12:18, 296 12:19, 154, 155, 291 12:21, 148 12:22, 151 12:23, 292 12:25, 294 12:27, 293 12:28, 289 12:31, 291, 294 12:32, 292 12:33, 291 12:34, 291 12:38, 153, 289 12:39, 153 12:40, 296 12:41, 290 13:2, 294 13:5, 291 13:7, 294 13:8, 291, 294 13:13, 294

315

Scriptural Passages 13:14, 289

14:4, 150 14:1, 151 14:3, 156 14:4, 152, 289 14:6, 289 14:7, 289 14:8, 150 14:10, 293, 294 14:11, 289, 294 14:14, 155, 294 14:15, 293 14:15–16, 155 14:16, 289, 293 15:11, 150 15:1, 153 15:2, 295 15:3, 158 15:8, 150 15:11, 289 15:12, 293 15:14, 156 15:16, 295 15:18, 148, 289 15:19, 155, 289 15:21, 290 15:23, 293 15:24, 148, 149, 160, 290, 295 15:25, 158 15:26, 152, 160 15:27, 152, 161, 295 15:28, 150, 290 15:29, 148, 295 16:1, 148 16:4, 153 16:5, 148, 290 16:6, 146, 158, 295 16:7, 292, 295 16:10, 158, 290 16:15, 292

16:16, 290 16:22, 291 16:23, 292 16:27, 291 16:30, 152 16:31, 152 16:35, 156 16:38, 290 16:40, 156 16:42, 157 16:43, 152 17:1, 148 17:4, 293 17:5, 290, 293 17:7, 157 17:9, 154 17:13, 153, 154, 161 17:16, 295 17:18, 152 17:19, 152 17:21, 153, 155, 156 17:23, 293 17:26, 291 17:27, 293, 295 18:3, 290 18:4, 157 18:5, 148 18:7, 295 18:11, 150, 152 18:12, 148, 160 18:16, 149, 295 19:1, 70 19:2, 290 19:3, 290 19:5, 154, 291 19:6, 290 19:7, 290 19:10, 149, 160 19:13, 294 19:14, 156

19:15, 293

19:16, 149, 153, 161, 162 19:17, 145, 292 2 Esd 9:9, 144, 145 13:15, 25 22:27, 25 Jdt 6:8, 69 16:12, 69 16:14, 69 16:18, 69 1 Macc 5:20, 22, 251 14:5, 252 4 Macc 2:7, 75 4:4, 96 4:21, 107 10:1, 106 11:1, 106 Job 35:5, 66 Ps 1:5, 168, 297 3:2, 168, 297 3:6, 182 5:10, 179 6:9, 175 7:7, 166 9:2, 175 9:24, 182 10:6, 170 12:4, 179 14:4, 166 15:1, 166 15:6, 191 15:9, 168, 175, 178, 298 16:2, 175 16:12, 170 17:6, 170 17:11, 166

316 17:19, 166 17:36, 166 18:7, 179 19:8, 191 20:5, 179 20:7, 183 20:8, 187 21:15, 191 22:1, 191 24:15, 170 25:2, 175 27:2, 179 27:3, 198 27:7, 172, 298 29:1, 187, 298 30:5, 171 30:21, 166, 191 30:23, 198 30:24, 166 30:25, 179 31:8, 189 33:7, 191 34:7, 171 34:8, 171 34:9, 166, 191 34:27, 191 36:5, 179 36:8, 191 36:10, 179 36:15, 191 36:23, 199, 200 36:27, 166 36:40, 175 38:13, 172 40:8, 189 40:10, 179 40:12, 187 41:5, 69, 75 41:6, 187 43:14–15, 176, 178 43:17, 191 43:23, 172 44:10, 185, 186

Scriptural Passages 44:14, 185 44:17, 189 44:18, 196 45:4, 183, 188 45:6, 189 47:9, 179 47:14, 176, 178 48:15, 192 48:18, 176 48:19, 168, 298 49:1, 171 50:20, 76 51:11, 186 54:7, 171 54:10, 192 54:21, 200 54:22, 200 55:5, 192 56:1, 179 56:2, 189 56:4, 189 56:7, 171, 183 58:2, 198 60:6, 189, 298 61:7, 198 61:9, 176 61:13, 92 62:11, 176 63:6, 171 64:5, 192 64:9, 192 64:10, 183, 192 65:5, 166 65:7, 179 65:8, 172 65:11, 171 65:13, 200 65:20, 183, 185 66:3, 183, 186 66:5, 186 67:5, 172, 298 67:17, 198 67:22, 192 67:28, 176, 178

67:31, 192, 298 67:32, 170, 196, 198 67:33, 196 67:34, 171, 198 67:36, 180 68:7, 192 68:12, 183 68:17, 172, 298 68:18, 189, 191, 298 68:20, 192 68:21, 170, 197, 198 68:23, 171 68:25, 200 68:31, 171 68:32, 198 68:35, 186 69:2, 192 70:2, 186 70:8, 166 70:10, 180 71:7, 168, 190, 298 71:10, 197, 198 71:19, 200 72:6, 197 72:20, 171 73:4, 188 73:6, 192 73:10, 198, 298 73:11, 176, 178 73:13, 192 75:12, 197 76:4, 75 76:5, 190, 299 76:9, 198 76:14, 198 76:16, 167 77:4, 173, 183, 299 77:5, 183 77:8, 75, 173, 299

77:15, 180 77:16, 171 77:19, 172 77:35, 198 77:40, 186 77:42, 175 77:48, 186 77:49, 168, 173, 299 77:69, 183 78:1, 168, 173, 174, 187, 299 78:6, 176, 192, 299 78:9, 180, 192, 299 78:11, 173, 299 80:1, 173 80:6, 190, 299 80:13, 173 80:14, 300 81:7, 192 82:6, 192 82:14, 183 82:18, 168, 176, 178, 300 83:4, 198 84:6, 190 84:14, 192 85:1, 192 85:4, 198 85:9, 190 85:14, 173, 192 86:4, 176 87:1, 198 87:7, 169, 300 87:15, 168, 173, 300 87:16, 197 88:4, 186 88:35, 192 89:2, 198 90:3, 171 90:12, 188

Scriptural Passages

317

91:4, 171 91:15, 180 93:14, 168, 173, 300 93:19, 183 94:4, 180 94:6, 193 95:7, 197, 198 96:1, 168, 300 96:5, 198 96:6, 180 96:7, 188 96:9, 187 97:3, 164 98:4, 176 100:3, 190 100:6, 193 101:12, 168, 184, 300 101:16, 199 101:20, 173 101:26, 180 102:2, 193 102:7, 180 102:15, 173, 193 103:3, 193 103:7, 180 103:8, 193 103:22, 193 103:25, 89 103:28, 193 104:1, 173, 300 104:14, 180 104:16, 193 104:20, 176, 178, 300 104:35, 184 104:39, 190 105:2, 190 105:7, 171, 301 105:10, 168, 173, 301 105:21, 193 105:23, 184

105:33, 180 105:34, 180 105:38, 173 105:46, 180 105:48, 167 106:4, 184 106:14, 188 106:30, 193 107:3, 193 107:10, 180 107:12, 193 108:4, 171, 301 108:12, 173, 301 108:14, 184 108:19, 193 108:21, 193 108:26, 187 108:29, 197 108:30, 180 109:3, 180 109:6, 171, 172, 193 110:6, 193 111:1, 193 111:4, 89 111:5, 184 111:7, 180 111:8, 176, 178, 301 112:1, 184 112:9, 184 113:1, 190 113:2, 180 113:4, 171 113:6, 171, 177 113:11, 193 113:23, 180 114:1, 193 114:2, 194 114:5, 199 115:3, 199 117:4, 177 117:5, 194 117:13, 181

318 118:1, 181 118:3, 169 118:15, 194 118:19, 188 118:20, 199 118:23, 167 118:34, 194 118:35, 194 118:36, 167, 194, 301 118:38, 177, 301 118:41, 184 118:49, 194 118:50, 177 118:51, 177 118:55, 194 118:56, 199 118:60, 187 118:63, 181 118:64, 199 118:69, 184, 301 118:70, 194, 302 118:72, 194 118:73, 189 118:75, 188, 197 118:77, 194 118:79, 181 118:82, 189, 302 118:88, 184, 302 118:92, 169, 302 118:95, 177 118:102, 190 118:109, 181, 194 118:112, 190 118:114, 194 118:115, 181 118:116, 181 118:122, 194 118:124, 184 118:127, 190, 302 118:134, 184, 302 118:141, 173, 302 118:147, 184 118:148, 173

Scriptural Passages 118:149, 169, 302 118:152, 199 118:163, 181, 182, 302 118:169, 181, 199 118:173, 181 118:174, 174 118:176, 171 120:6, 167 120:4, 194 120:6, 194 120:8, 190, 303 121:4, 199 121:6, 197 121:7, 181, 190 121:9, 199 122:2, 75 122:3, 194 122:4, 194 123:7, 174 123:8, 181 124:4, 184, 303 126:2, 174 126:3, 181 127:3, 174, 303 127:4, 181 127:6, 194 129:5, 167 129:6, 177 129:8, 174 130:1, 177, 178, 197, 303 130:2, 177, 181, 194, 303 103:12, 188, 300 131:5, 174, 303 131:7, 187 131:13, 195, 303 131:17, 76 132:2, 190 133:1, 177, 178, 303 133:2, 181 134:6, 174, 303

134:7, 181 134:8, 174, 304 134:10, 195, 304 134:18, 195 135:2, 64 135:3, 64 135:11, 181 135:14, 195 135:23, 177 135:24, 181 136:3, 195 137:3, 167 137:5, 197 137:7, 190 138:5, 190 138:6, 190, 304 138:9, 195 138:18, 181 138:19, 190 138:23, 178, 179 139:2, 174, 304 139:5, 199 139:11, 197 139:12, 185 139:14, 185 140:1, 181, 185 140:4, 195 140:6, 182 140:7, 185 142:2, 195, 304 142:8, 182 143:2, 182, 195 143:3, 195 143:11, 199 143:13, 174 143:15, 195 144:1, 195 144:4, 174 144:6, 178 144:10, 195 144:13, 174, 304 144:17, 185 144:19, 195 144:20, 185

Scriptural Passages 145:4, 195 145:5, 195 146:1, 182 147:1, 168, 182, 304 147:4, 182, 195 147:5, 190 147:7, 195 148, 38 148:1, 174, 304 148:10, 185 149:1, 199 149:6, 182 149:9, 182, 185, 197 150, 38 150:5, 174 151, 38 Eccl 4:9, 120 Isa 1:1, 100 1:4, 101 1:21, 101 1:26, 101 3:3, 101 3:6, 66 3:20–21, 102 7:22, 88 24:7, 17 26:15, 88 28:14, 66 28:22, 17 29:11, 17 50:3, 87 63:16, 66 64:11, 66 Jer 3:18, 66 3:25, 66 4:7, 22 5:23, 88 31: 33–34, 154 31:34, 212

Joel 1:2, 81 2:20, 96 4:6, 66 Jonah 1:3, 92 1:14, 88 Hab 3:16, 85 Matt 1:1, 283 1:15, 86 1:17, 283 1:18, 96, 114, 263 1:24, 222 2:5, 264 3:12, 283 4:3, 264 4:11, 264 4:21, 222 5:45, 12 7:21, 283 7:26, 102, 103 9:36, 283 9:37, 283 10:8, 264 10:17, 113, 283 11:10, 264 11:13, 264 11:24–25, 103 11:27, 265 12:8, 265 12:33, 48 12:44, 85 13:1, 283 13:7, 230, 232 13:24, 265 13:35, 265 13:54, 19, 20, 252 13:57, 115, 265 14:24, 266 14:35, 104 15:24, 117, 283 15:29, 104

319 16:2–3, 117 16:2–6, 120, 266 16:4, 117 16:6, 284 16:7, 284 16:9–18:12, 40 19:15, 266 19:21, 284 19:28, 284 21:9, 65 21:21, 284 21:22, 284 22:9, 48 22:16, 266, 284 23:13, 65 23:30, 140 23:32, 267 24:16, 115, 267 24:23, 284 24:30, 267 24:36–25:21, 40 25:18–26, 69 26:11, 267 26:21, 267 26:45, 267 27:2, 268 27:11, 268 27:16, 284 28:1, 284 28:4–5, 268 Mark 1:3, 284 1:4-10, 268 1:9, 268 1:11, 269 2:16, 269 3:24, 269 4:2, 115, 270 6:6, 270 6:15, 270 6:21, 270 6:31, 236 6:34, 271 6:47, 281

320 7:6, 115, 271 7:20, 65 7:26, 271 7:33, 272, 284 7:36, 272 8:1, 272, 284 8:12, 281 8:22, 281 9:33, 113, 114, 281 9:43, 118, 272 9:48, 281 10:17, 115, 273 10:21, 112 10:32, 273 10:35, 281 10:41, 114 10:46, 115, 273 11:19, 281 11:27, 273 11:41, 117 13:1, 284 13:3, 104, 106, 112, 273 13:17, 284 14:7, 273 14:18, 274 14:32, 274 14:41, 274 14.54–16:8, 40 14:54, 284 14:69, 274 15:15, 67 15:19, 274 15:23, 281 15:25, 281 15:28, 118, 119, 120 15:29, 281 16:3, 275 16:6, 69 Luke 1:1, 223 1:1–56, 40

Scriptural Passages 1:4, 223 1:5, 94 1:13, 229, 231 1:26, 234, 235 1:39, 234, 235 1:49, 229, 231 1:65, 232 1:69, 225, 227 2:1–2, 105 2:2, 232 2:3, 240 2:4, 228 2:6, 231 2:7, 234, 235 2:9, 234, 235 2:12, 225, 234 2:15, 228 2:21, 230, 231 2:22, 233 2:27, 237 2:28, 233 2:29, 225 2:36, 229, 231 2:37, 234, 236, 237, 238 2:39, 228, 234, 235 2:41, 229, 231 2:42, 225, 228 2:46, 223 2:52, 230, 231 3:1, 226, 237 3:5, 234 3:7, 233 3:15, 221, 222 3:16, 228 3:17, 221, 222, 229, 231 3:18, 226 3:21, 119, 284 3:23, 284 3:29, 223 3:32, 223 4:1, 233, 237, 238

4:2, 233, 245 4:4, 223, 224 4:5, 222 4:36, 222 4:38, 228 4:41, 229, 238 4:42, 237 4:43, 222 5:1, 237 5:2, 237, 238 5:3, 223 5:5, 236 5:7, 234 5:13, 226, 234 5:14, 237, 239 5:18, 230, 232 5:32, 234, 235 5:36, 228 6:2, 223, 224, 225 6:14, 237 6:17, 224, 225 6:26, 236 6:31–32, 113 6:36, 115, 275 6:44, 234, 235 6:47, 228 7:2, 275 7:3, 228 7:6, 226 7:7, 226 7:11, 237 7:17, 284 7:27, 275 7:28, 233 7:35, 230, 232 7:36, 229, 231, 284 8:6, 228 8:7, 230 8:10, 233 8:11, 234, 235 8:13, 228, 229, 231 8:15, 226

Scriptural Passages 8:18, 275 8:22, 228, 284 8:26, 234, 235 8:29, 228, 230, 231 8:30, 237 8:36, 237, 238 8:40, 234, 236, 281 8:43, 226, 228 8:47, 237, 239, 240 8:55, 237, 239 9:3, 224, 225 9:7, 237 9:9, 224 9:13, 234 9:21, 102, 226, 227 9:28, 236 9:32, 226 9:39, 226, 227 9:41, 236 9:43, 237, 238 9:51, 228 9:62, 224, 225 10:11, 228 10:13, 103 10:15, 233 10:20, 226 10:30, 230, 231 10:42, 237, 239 11:1, 238 11:2, 238 11:18, 234, 235 11:22, 228 11:27, 228 11:28, 230 11:36, 226 11:40, 228, 229 11:42, 226 11:51, 224 12:4, 226 12:8, 238

12:14, 235 12:18, 235 12:26, 233 12:29, 230, 232 12:32, 236 12:42, 235 12:52, 227 12:57, 89 13:14, 225 19:16, 96 21:24, 110, 115 21:34, 281 22:40, 24 John 1:14, 85 1:30, 284 1:41, 276 3:14, 72 3:36, 281 6:30, 116, 284 6:38, 114, 116, 285 6:49, 113 6:61, 276, 285 6:65, 281 7:28, 285 7:33, 285 7:41, 276 8:20, 285 8:38, 25 10:7, 104 11:54, 276 12:24, 285 12:27, 115, 277 13:13, 285 13:16, 119, 285 13:21, 285 13:23, 113, 116, 277, 281, 285 13:26, 17 13:31, 72 14:15, 277, 285 14:21, 277 14:25, 115, 277

321 16:14-16, 278 18:1, 285 18:12, 119, 285 18:16, 278, 281 18:20, 285 18:28, 278 18:37, 278, 285 18:38, 285 18:39, 285 19:2, 279 19:5, 285 19:8, 285 19:14, 279 19:20, 67 19:21, 116, 285 19:23, 285 19:25, 115, 279 19:28, 286 19:31, 279 21:1, 119, 286 21:12, 113, 280 21:15, 280 Acts 3:1, 122 5:1, 122 6:9, 122 8:5, 19, 252 8:9, 123 8:26, 123 8:40, 19 9:22, 123 9:32, 123 9:36, 123 10:1, 123 13:6, 123 15:40, 121 16:13, 123 17:16, 123 18:1, 123 18:15, 87 18:24, 123 19:13, 123 19:24, 123 20:7, 123

322

Scriptural Passages

20:17, 123 21:15, 123 21:26, 87, 124 21:39, 124 22:6, 124 23:26, 124 23:31–33, 253 24:10, 124 25:9, 124 26:1, 124 27:1, 124 Rom 1:8, 217 1:12, 205, 206 1:16, 205 1:17, 89 1:28, 203 2:16, 217 2:25, 204 3:10, 89 3:26, 89 3:31, 205 4:2, 203 4:12, 206, 207, 217, 218 4:13, 208 4:18, 214 5:1, 203 5:3, 204 5:5, 215 5:7, 89, 204 5:9, 214 5:16, 97, 107, 205 5:17, 107 6:22, 214 7:1, 108 7:4, 64 7:7, 108 7:15, 204 7:25, 204 8:5, 97, 107, 108 8:7, 205 8:30, 89 9:3, 216

9:12, 204 9:27, 204 9:32, 210, 212 10:1, 205 10:5, 216, 217 10:16, 208, 209 10:17, 204 11:1, 204 11:8, 214 11:22, 210, 212 11:30, 217 11:31, 209 14:20, 211, 212 14:23, 207 15:2, 217 15:9, 211, 213 15:12, 205, 206 15:28, 210, 212 15:32, 216 16:2, 213 16:7, 209 16:25, 211, 213 1 Cor 1:1, 64 4:4, 89 14:1, 65 Col 1:7, 204 1:19, 215 1:27, 215 2:1, 217 2:3, 205 2:7, 217 2:8, 205 2:10, 218, 219 2:14, 202, 205 2:18, 207, 209, 215 2:20, 216 2:23, 205 3:1, 209, 211, 213, 215 3:10, 215 4:2, 217

4:3, 210, 212 4:7, 217 4:8, 205 4:11, 69 4:12, 215 1 Thess 1:5, 213 1:8, 204 2:9, 215 2:13, 22, 45, 215, 217 2:14, 40, 207 2:14–5:28, 40 3:11, 214 4:8, 70 4:10, 214, 216 4:14, 215, 219 5:1, 208, 209 5:8, 217 2 Thess 1:3, 209 1:7, 215 2:4, 217 2:7, 214 2:14, 215 2:16, 209, 217 3:8, 204 3:10, 209 Titus 1:16, 48 Heb 1:1, 205 1:5, 209 1:8, 207 1:12, 209 2:4, 214 2:7, 206 2:8, 215 2:18, 217 3:6, 215 3:8, 210, 212 4:7, 216 4:9, 218 4:11, 209

323

Scriptural Passages 4:12, 218 4:14, 214 4:16–8:1, 41 7:10, 217, 218 7:21, 217, 218 8, 154 8:8, 214 8:10, 215 8:12, 212 9:12, 207 10:7, 218 10:11, 216 10:12, 213 10:17, 212 10:18, 215 10:26, 210, 212 10:32, 211, 212

10:34, 206 10:36, 214, 216 10:39, 214 11:3, 206 11:5, 215 11:8, 218 11:9, 209 11:11, 89 11:22, 66 12:4, 206 12:10, 209, 211, 212 12:20, 211, 213 12:21, 218 12:23, 218 13:2, 215 13:12, 214

13:18, 206 13:22, 209 13:23, 215 1 John 2:24–5:5, 97 Rev 1:8, 67 2:1, 98 2:8, 98 4:8, 98 11:8, 67 17:14, 64 19:16, 64 21:6, 67 22:13, 67 22:20, 64